Skip to main content

Full text of "Works"

See other formats
























All rights reserved 




The present volume consists of selections from Professor 
Green's unpublished philosophical papers. It was his 
practice, both as college-tutor and as professor, to 
write out and keep full notes for most of his lectures. 
These were rewritten and amplified from time to time, 
and in some cases developed into tolerably finished 
compositions. In making selections from them it has 
been thought advisable not to include anything written 
before 1874, the date of the ' Introductions to Hume ' 
(see vol. I.) The earlier drafts, though by no means 
devoid of interest, are for the most part superseded by 
those which are here printed ; and where this is not the 
case, the more careful composition of the latter seems to 
show that they contained the writer's maturer views. 

Though not intended for publication, the manu- 
scripts were in general continuous and coherent, and with 
a few unimportant exceptions they have been printed 
without change of form or expression. In cases where 
the order or connexion of passages was not obvious, 
I have had to exercise my discretion. I am also respon- 


sible for the division into sections, the table of contents, 
and the notes and insertions in brackets. 

My best thanks are due to Professor E. Caird for 
advice in selecting the manuscripts for publication, to 
Professor A. C. Bradley and Mr. J. C. Wilson for help 
in arranging and preparing them for the press, to Mrs. 
Green for copying a great part of them, and to her and 
Miss Green for reading the proof-sheets. 

Oxford : February, 1886. 





Lectures on the Philosophy op Kant. 
I. The 'Critique op Pure Season.' 

A. Kant's problem, and the relation of the statement of it to 
the philosophies of Locke and Hume, 

1. The three dominant ideas in Kant's speculation come to him 

through Locke and Hume ...... 2 

2. Difficulties arising from Locke's opposition between mathe- 

matical and physical truth ....... 2 

3. Hume solves them by resolving them both into l impressions ' . 4 

4. False antithesis between l object' and 'thought,' resulting 

in that between 'particular' (contingent) and 'universal,' 
(necessary) propositions 5 

B. The ' forms of perception ' (' intuition '). 

5. Inconsistent views of ' object ' in Kant, as (1) given independ- 

ently of thought, (2) constituted by thought acting on 
matter of sensation ........ 8 

6. Space and time are both relations of objects and ideas, dis- 

tinguished by primariness and simplicity .... 9 

7. Kant's distinction between the idea of space as ' intuition' from 

other ideas as ■ conceptions ' . . . ... 1 

8. It gives no ground for opposing ' intuition ' as what is ' real ' 

to * conception ' as the ■ work of the mind ' . . .11 

C. The ' deduction of the categories ' in the first edition of 
the ' Critique* 

9. Point of Kant's question, How can pure conceptions relate 

a priori to objects ? . . . . . . . . 13 

10. The ground for assuming such conceptions is that related 

phenomena imply a correlating unit . . . . . 11 



11. And we find no such unit except in consciousness . . . 15 

12. Kant's account of this unity of consciousness. (1) ' Syn- 

thesis of apprehension ' . . . . . . .16 

1 3. Particular difficulties apart, Kant's point is that a synthesis 

is necessary to constitute an individual object, this or that. 17 

14. ' Synthesis of apprehension ' involves (2) that of ' reproduc- 

tion.' Relation of Kant's doctrine to that of ' association ' . 19 

15. (3) * Synthesis of recognition in the concept.' Kant's un- 

tenable opposition between * intuition ' and ' conception/ 
involving an ambiguous use of ' object.' .... 21 

16. This conceptual synthesis is implied in the ordinary account 

of simple perception . . . . . . . . 22 

17. It is also, as ' transcendental apperception,' the condition of 

there being ' reality ' or a ' real object ' . . . .23 

18. Its application by Kant to explain the meaning of ' the corre- 

spondence of thought with its object' 24 

19. As that without which there would be no ' objects ' at all, it 

is ' transcendental ' . . . . . . . .25 

20. It is not dispensed with by saying that ' object ' = merely the 

'possibility of experience,' for the 'possibility' must be a 
determinate possibility . . . . . . . 27 

21. Thus what Kant calls the 'transcendental object ' and the 

1 transcendental subject ' are the same .... 27 
22 Is it the ' thing in itself ? Divergence as to this between the 

Critique, Ed. I and the Prolegomena . . . . . 28 

23. Kant writes wrongly as if ■ intuition ' could give ' objects ' 

independently of ' conception ' 29 

24. Which suggests that f phenomena' or affections of sensibility 

are due to some other cause than the unity of self-con- 
sciousness 30 

25. If this unity 'makes nature,' must not either (1) nature be 

'subjective,' or (2) myself =God? Kantian answer to (2). 30 

26. As regards (1), it is only by false abstraction that we can 

separate our successive states of mind from eternal rela- 
tions 31 

27. Kant's ' deduction ' in Ed. I really answers the qucestio facti 

as well as the qumstio juris as regards the ' pure concep- 
tions ' 32 

D. The ' schematism of the pure conceptions of the under- 
standing^ and the ' mathematical principles' 

28. 'Schemata' are determinations of internal sense which medi- 

ate between (1) categories and objects of intuition, (2) one 
conceived object and another . . . ... 35 



29. The partial error in Kant's account of them is due to his 

false separation of categories from ' phenomena,' as if the 
latter were objects without the former .... 36 

30. Properly speaking no objects are ' given ' without ' unity of 

apperception ' (of which the categories are forms), and 
therefore no mediation between them is needed . . . 37 

31. Kant's account of number as schema of quantity involves 

confusion between time as condition of counting and time 

as what is counted ........ 39 

32. The apprehension o£ an object of intuition ipso facto consti- 

tutes it a quantity ........ 41 

33. Thus number does not render apprehension of extensive quan- 

tities possible; it is itself the result of the apprehension in 

its most abstract form . . . . . 42 

34. Kant's account of intensive quantity as the ' real ' in phe- 

nomena .......... 43 

35. He uses ' apprehension ' both for consciousness of sensation 

and for consciousness of it as the real in phenomena, i.e. 

as degree . . . . .... 44 

36. Want of adjustment between the 'real' as 'schema,' (time- 

filling), and the ' real ' in the ' anticipations of perception,' 
(degree) 45 

E. The relation of the ' mathematical 1 to the ' dynamical } 

37. According to Kant, the mathematical principles represent 

' constitutive,' the dynamical ' regulative,' knowledge . 46 

38. But (a) the relations in which a phenomenon must stand to 

another 'concern' the phenomenon, and a phenomenon 
cannot be intuited or perceived without such relations . 47 

39. And (li) we do not ' anticipate experience ' in the ' axioms ' 

and ' anticipations ' any more than in the ' analogies ' . 48 

40. Nor is the necessity of the mathematical principles rightly 

contrasted with that of the dynamical as ' intuitive ' with 

' discursive ' 48 

F. The proofs of the ' analogies of experience. 1 

41. According to Kant, the three 'analogies' correspond to the 

three l 7yodi of time 1 50 

42. They represent relations of 'time in general ' (as opposed to 

the time at which phenomena happen to occur) in which 

the thinking subject asserts its presence . . . . 51 


43. (1) Proof of the first 'analogy.' Kant is right in saying 

that no relation of time is possible without permanence, 
wrong in calling permanence 'another expression for 
time' 52 

44. Time, as the relation of succession, is a permanent relation, 

but it is not itself ' the permanent/ which is just the 
negation of succession 53 

45. Time being either succession or duration, succession, per- 

manence, and coexistence are not rightly called ' modes of 
time' 54 

46. The only real permanent, which makes all modes of time pos- 

sible (Kant's * matter of all phenomena '), is the eternal 
subject .......... 54 

47. The second ' analogy.' The relation of cause and effect 

implies a distinction between the subjective sequence of 
feelings and the objective sequence of events, and deter- 
mination of the former by the latter . . . .55 

48. The source of such determination is the unity of understand- 

ing, in virtue of which the order of events is one . . . 57 

G. The distinction of ' analyticaV and ' synthetical 'judgments 
in connexion with the 'postulates of empirical thought.'' 

49. Kant's doctrine that 'possibility of experience ' gives 'objec- 

tive reality to a priori cognition,' as giving the principle of 

all synthetical judgments 58 

50. The principle of contradiction is not rightly called the 

' test ' of the truth of analytical judgments, for either they 
need no test, or, if they do, they are not analytical . . 59 

51. Kant's distinction between ' all bodies are extended ' and 'all 

bodies have weight ' is untenable : both involve ' experi- 
ence' . . . . . . . . . .61 

52. Both imply synthesis and both analysis. Kant really supplies 

the correction to his own doctrine 62 

53. His account of the ' principles of modality ' partly suffers 

from the same notion, that the conceived object can be 
thought apart from its relations 62 

54. A 'possible' object is either 'objectively possible,' in which 

case it is real, or ' subjectively possible,' in which case it 
cannot be conceived as real without ' further determina- 
tion' ... 64 


H. The distinction between i outer' and ' inner' sense. 


55. The true sense of this distinction is that between the con- 

sciousness of objects as related to each other and as changes 

in the conscious subject ....... 65 

56. With Kant 'internal sense' generally = * empirical apper- 

ception,' which implies a ' permanent external something ' 

or ' substantia phenomenon ' . • . . . . 66 

57. He is right in holding that consciousness of our successive 

states as ours implies that of permanent objects other than 
them .......... 67 

58. But (1) such consciousness is not rightly described as ' out- 

ward/ and (2) of ' inner sense ' properly so called time is 

not the peculiar form . . . . . . . 67 

59. Kant uses ' internal sense ' for (1) the sensible in general, of 

which time is the form, (2) empirical apperception ; and in 
both it implies action of the understanding ... 68 

60. Recapitulation of the meanings of the distinction between 

' outer ' and ' inner ' sense . . . . . . . 70 

I. The l empirical reality ' of time. 

61. Kant's doctrine does not at first sight seem to allow enough 

reality to space and time . ... . . .72 

62. Mere time is an abstraction ; in the real world of becoming 

time exists only as a factor neutralised by what is not time. 
Three questions as regards such a world . . . . 72 

63. The events which preceded the appearance of sentient life 

form one whole with our experience, and the condition of 

this whole is not our sentient life, but an eternal reason . 73 

64. As being ourselves this reason we conceive nature as a 

1 system of becoming ' ; as being ourselves part of nature, 

we do not get beyond this formal conception . . . 74 

65. As such a system, nature must be conceived as both finite 

and infinite ......... 75 

66. If with Kant we say that this contradiction is inherent in 

1 empirical reality,' which =. sensibility determined by 
reason, what becomes of ' reality' before human sensibility 
existed? . . . 76 

67. Of three alternative answers two untenable ones may be ex- 

tracted from Kant's language 77 

68. The only tenable view is that nature, as the eternal fact of 

change, implies an eternal sensibility, determined by reason, 

and reproduced in up . . . . ., . . . 79 



69. Answers to questions raised in sec. 62 ; (1) nature exists only 
for a thinking subject; (2) for such a subject it is both 
finite and infinite ; (3) for such a subject time exists, not 
in itself, but as a factor in change 80 

II. The Metaphtsic op Ethics. 

K. The distinction between ' natural * and i moral ' agency. 

70. Is moral philosophy a branch of anthropology ? . . .83 

71. Or is moral agency absolutely distinct from natural, as held 

by Kant ? . . . . 83 

72. Necessity of starting from Kant, in distinguishing moral 

philosophy from natural science 84 

73. Kant's dictum that ' understanding makes nature.' Its gene- 

ral meaning . . . . . . . . . 85 

74. Difficulties in Kant's account of the relation of understanding 

to sensibility ......... 86 

75. Problems to which he conceives that the process of know- 

ledge gives rise ......... 87 

76. What he distinguishes as * reason ' and ' understanding' are 

really forms of the same principle ..... 88 

77. Sense in which he may be said to hold that ' knowledge is 

only of phenomena ' . . . . . . 89 

78. The function of reason in relation to desire as distinguished 

from its function in knowledge. 'Want' and 'wanted 
objects' 90 

79. The self- consciousness involved in wanted objects is not 

strictly ' natural ' . . 91 

80. But if ' nature ' also implies self- consciousness, why distin- 

guish ' moral ' from ' natural ' ? Different senses in which 
understanding ' makes ' nature and ' makes ' morality . 92 

81. In nature it is not itself an agent in the series of events which 

it constitutes, as it is in moral action . . . . . 93 

82. As thus directly affected by the reason of the individual 

agent, moral action is ' free ' as no natural action is . . 94 

83. But ' free ' in this sense does not mean ' unmotived' . . 95 

84. Thus we may adopt Kant's distinction between a * natural ' 

and a ' rational ' agent . . . . . . .95 

L. The distinction between the l empirical ' and l intelligible * 

85. Kant's doctrine of ' free causality.' The 3rd ' antinomy ' 

(1) in its cosmological bearing . . . . . . 98 



86. (2) In its ethical bearing. Apparent unsatisfactoriness of 

the distinction between the ' empirical ' and ' intelligible ' 
characters ■.*..'. . . . . . 100 

87. Kant himself supplies the means for getting rid of it . . 101 

88. Recapitulation of criticism on Kant, and questions to which 

his doctrine gives rise . . . . . . .104 

M. The coexistence of freedom and natural necessity in the same 


89. Kant's moral theory forces upon him the problem, how free- 

dom and natural necessity can coexist in the same action . 106 

90. Are some acts free, others naturally determined ? or is the 

same act jointly determined ? . . . . . .107 

91. Kant cannot admit either alternative. (His double use of 

1 freedom ') 107 

92. The truth is that the real man can only exist and act in one 

way, i.e. as an object to himself, and in that sense ' free' . 108 

93. The right and wrong in Kant's account of the matter . . 109 

N. The good will. 

94. The problem of idealistic ethics is to justify the conception 

of that which unconditionally ought to be . . . .110 

95. The essence of Kant's solution . . . . . . 110 

96. What Kant calls Maw ' really implies 'object' . . . Ill 

97. The primary question which his doctrine raises is, Can reason 

constitute a law or object ? His use of ' reason ' . . . Ill 

98. His conception of the function of unity in knowledge, in its 

two forms of ' reason ' and ' understanding ' . . .113 

99. Connexion between the \ reason ' of the Critique of pure 

reason and that of the ethical treatises. . . . . 114 

100. Difference between them . . . . . . .116 

101. Two ways in which the former prepares the way for the 

latter . . . . . 116 

102. Meaning of Kant's question, 'Is there a practical use of 

pure reason ? ' or, 'Is there a free causality ? ' . .117 

103. It. is answered by an appeal to the 'voice of conscience.' 

Relation of the Grundlegung to the Kritik d. prakt, 
Vetyiunft . . . . . . . . . . 119 

104. Kant's assertion that the idea of duty is no 'Erfahrungs- 

begriff.' Sense in which he was right . . . .121 

105. How far was he right in calling the a priori element in 

moral experience an ' idea of universal law ' ? . . 124 

106. Nothing is really gained by ascribing it to ' reason ' ; the 

question is, Is the analysis which yields it true ? . . 125 



107. Explanation and connexion of Kant's two first formulations 

of the * categorical imperative ' 126 

108. The third formulation of it 128 

109. The basis of his ethics is that ' good will ' = ' autonomous 

will/ Three questions raised by this doctrine . . . 130 

110. (1) Is there such a thing as determination of will by law 

in virtue of its mere form ? Is not such law objectless? . 130 

111. No; Kant's conception of such law does imply relation to 

an object, i.e. the desirable before it is desired . . . 131 

112. Which he finds in the rational nature of man. But this 

again raises several questions 132 

113. Is not Kant's 'reason' merely the' animal instinct of self- 

preservation ? No ; for instinct excludes self-conscious- 
ness 133 

114. Why ascribe this self-consciousness to ' reason' ? Because 

it is the same principle that is involved in knowledge . 133 

115. This principle in activity is * will.' Kant's three defini- 

tions of will ......... 134 

116. Essential difference between them and the definition e.g. of 

Bain 135 

117. < Kational ' in the third definition must be interpreted so as 

to include \ heteronomous ' will 136 

118. Does then Kant's theory give a ground for distinguishing 

f good ' from ' bad ■ will, as embodied in various concrete 
interests? 138 

119. Not in the sense that all interests are either interests in the 

fulfilment of the moral law or desire for pleasure : some 

are neither 132 

120. Does then Kant's principle, that only the interest constituted 

by reason is truly moral, furnish a criterion of moral 
value? 140 

121. We must bear in mind that the only subject of moral valua- 

tion is character. Difference between ' desiring pleasure ' 

and ' living for pleasure ' 141 

122. Character means the way in which a man seeks self-satis- 

faction . . . . 142 

123. ' Pleasure-seeking ' is not the only immoral, nor consciously 

1 living for the good of others ' the only moral, life . . 144 

124. Justification of Kant's account of the 'good' will, as 

regards (1) its object, (2) its binding character . . 145 
3 25. Kant's doctrine that the moral law is ' self-imposed.' 

(1) In what sense is it true ? . • . . . . 146 

126. (2) How far is the consciousness of its being self-imposed 

a necessary condition of virtue ? 148 


O. Impossibility that desire for pleasure should yield a 
moral law. 


127. Desire for pleasure, even if uniform, could only give rise to 

a ' natural ' law ........ 151 

128. Distinction between ' physical ' and * moral ' necessity . 152 

P. The chief points of difficulty in Kant's moral theory. 

129. Doctrines, (1) that the idea of the moral law is not derived 

from experience ; (2) that the moral goodness of an act 
is not affected by any result ; (3) that the morally good 
act must be devoid of any motive ; (4) that the moral 
law is ' objectively necessary ' 154 

Lectuees on Logic. 
I. The logic of the formal logicians. 

1 . Different views of logic. (1 ) View that it is the science of the 

method of knowledge, modified according to the view taken 

of the object of knowledge 158 

2. (2) View of the formal logicians derived from the Aristo- 

telian and scholastic doctrine of syllogism ; this has only 

a practical value 159 

3. But is made by modern logicians into a speculative science, 

investigating the laws of formal thinking . . . . 161 

4. But (a) ' formal thinking ' is not a real process of thought 

at all 162 

5. As is practically admitted when it is expressed in a quantified 

form. 163 

6. Does it then (b), as described by formal logicians, represent 

the completed result of thinking ? . . . . .164 

7. * Intuition ' and ' sensation ' as used by Mansel and Hamilton 165 

8. Intuition cannot be opposed to conception as ' presentative ' 

to ' representative,' for the most immediate presentation 
implies something not present . . . . . . 167 

9. And an unconceived object of intuition would be a nonentity 

(devoid of attributes) . ....... 169 

10. For 'here' and 'now' express relations only possible to a 

subject not 'here ' and ' now,' i.e. a thinking subject . . 170 

11. And the felt thing (as distinct from feeling) is the ideal thing, 

a centre of intelligible relations . . . . .171 



12. What then is the relation between (a) intuition and concep- 

tion ? According to Kant, ■ thought ' is possible without 
intuition, but not ' knowledge ' 172 

13. Mansel's view seems to confuse the logical individualisation 

implied in all conception, with individualisation in space 

and time (intuition) . . . . . . 178 

14. There are many conceptions (e.g. moral ones) which are 

neither realisable nor verifiable by intuition. (' Faith ') . 174 

15. (b) As to the antithesis between perception and imagination, 

it is not that between (i) * work of things outside us ' and 

* work of the mind,' or (ii) ' real ' and * unreal ' . . 176 

16. (c) As to that between the real thing and the conception of 

it, (i) if the ' real ' be identified with feeling, it can only be 

so as related feeling 177 

17. (ii) If it be identified with 'matter,' it can only be matter as 

potentially related to all things, not the ■ matter ' of physics 177 

18. Thus without thought not only would there be no conscious- 

ness of relation, there would be no relations . . .179 

19. The permanence of the psychical effects of c registered ' feel- 

ings is quite different from their permanence as felt facts, 
which implies that they pass and are yet retained as past . 180 

20. The doctrine that thought constitutes relations does not (a) 

4 do away with external matter,' nor (b) resolve everything 

into thought 131 

21. Nor (c) imply that the universe is the creation of l my mind ' 182 

22. Any other doctrine obliges us to suppose (a) that the world 

which each man knows begins and ends with his birth and 
death . . . 183 

23. Or (b) that the^ is a world of ' things in themselves,' to 

which our c?' ;gories do not apply. (Sense in which we 
may speak 1 1 a spiritual world) . . . ... 183 

24. Returning .6, it may be said (1) ' the real thing is sensible, 

the conception is not.' But the real thing is sensible only 
because conceived. The conception differs from it through 
detachment (a) from sensation, (b) from other relations . ' 184 

25. A perfect conception of a thing is not distinguishable from 

the possible sensible experience of it, and to a perfect 
intelligence there would be no difference between possible 
and actual sensation • •.... 185 

26. (2) * We can make conceptions, not real things.' But if what 

we can make does not amount to reality, no more does 
what we cannot make (sensations) Igg 

27. (3) ' The real thing is individual, the conception is represen- 

tative.' But individuality, in whatever sense, implies or 
consists of conceptual relations .... 187 



28. The real, then, is real as conceived, and the relations to sense, 

which as perceived it implies, are conceived relations . 188 

29. All our conceptions of things are inadequate, and in process 

of formation, and therefore strictly speaking indefinable . 189 

30. To a perfect mind thought (not feeling) is the reality, and to 

us feeling does not add anything to thought except occa- 
sions for further thought . . . . . .190 

31. Incidentally it has been shown that the function of thought 

is not (according to the old notion) to abstract attributes 
from objects given independently of it . . 192 

32. The real process of all thought is from the abstract to the 

concrete, a joint process of analysis and synthesis . . 193 

33. The great distinction is between reflective and unreflective 

stages of thought . . . . . . . . 194 

II. The Logic op J. S. Mill. 

A. The import of propositions. 

34. Mill's order of inquiry (names, nameable things, propositions) 

is misleading . . . . . . . . .195 

35. The right order would be to begin with the analysis of pro- 

positions, which would yield the primary conceptions pre- 
supposed by all experience . . . - . . . . 196 

36. According to Mill propositions relate to ' things,' not ' ideas,' 

as with Locke. But he does not escape Lbcke's difficulty 

by substituting ' phenomenon ' for ' idea' . . ' . 197 

37. In Locke's view Mill's propositions about phenomena could 

only express particular and contingent tru k . . . 198 

38. If phenomenon ' be taken to mean ' appearand/ the difficulty 

remains how ' appearances ' come to be interpreted into an 
order of nature . . . . . Ijol . .199 

39. Mill, while rightly referring propositions to the real, does 

not see that the real implies a conceptual construction, the 
relation of thing and quality . . . . . . 200 

B. Names. 

40. Mill's distinction of ' singular ' and ' general ' names is really 

one of propositions ........ 202 

41. And ignores the different senses of ' individuality,' none of 

which can be expressed without a proposition . . . 202 

42. Singular names imply singular propositions, and these imply 

a conceived object of actual or possible intuition . . 203 
vol. ii. a 




43. Thus general propositions, though, if they concern nature, 

they must be verifiable in singular ones, express relations 

at once real and universal 204 

44. Mill's 'non-connotative' names are really connotative. Those 

which * signify attributes only ' really express substantiation 

of the attributes « • • 205 

45. ' Positive' and ' negative,' into which he divides names, are 

really correlative terms applying to every judgment . . 206 

O. Categories. 

46. Two views of categories, (1) as primary relations condition- 

ing experience, (2.) as ultimate things abstracted from 
experience ......-•• 

47. (a) l States of consciousness.' The knowledge of these as 

* existences ' already involves the category of substance and 
attribute 208 

48. And the distinction of them from their objects and from pro- 

perties of these involves the category of cause and effect . 208 

49. (b) l Mind ' and ' body.' These represent the same two 

categories, substance and cause, in abstraction from their 
correlative factors, attribute and effect . . . . 209 

50. Mill confuses body and mind partly with the outward cause 

of sensation and the sensitive organism, partly with the 

' thing in itself ' 210 

51. Thus 'states of consciousness' (1) are wrongly co-ordinated 

with substance and attribute, and (2) do not come under 
primary conceptions, which categories should be . . e 211 

52. If categories be understood as ' classification of existences,' 

no such classification as would include 'states of conscious- 
ness' could be exhaustive . . . . . .211 

53. (c) ' Attributes.' As regards quality and quantity Mill 

ignores the distinction between feeling proper and felt 
thing or fact . 212 

54. And does not recognise that judgments respecting them in- 

volve relations no less than those respecting what he calls 
specifically ' relations ' . . 213 

55. Relations which the related objects jointly produce only differ 

in the complexity of the relationship from any other 
qualities .......... 214 

56. Relations of resemblance and succession cannot be reduced 

(as Mill tries to do) to compounds of feelings . . . 215 

57. He does not see that all judgments involve the relation of 

identity, i.e. something other than mere feelings . . . 215 



58. Judgment is the thought of an object under relations, of 

which identity, cause, substance, reciprocal determination, 

are the most primary . . . . . . .216 

59. Resemblance, simultaneity, and succession are not rightly 

classed among relations involved in all judgment . . . 217 
GO. Mill holds two inconsistent views of the proposition, as 
expressing relations (1) between ' attributes,' (2) between 
* phenomena,' i.e. feelings . . . . . .218 

61. No judgment expresses relations of feelings as felt, and the 

relations which it does express are not merely those of 
feelings to one another, but of attributes to a thing . . 218 

62. Thus a judgment may state the sequence or coexistence of 

phenomena to be an attribute, but it never states the 
sequence or coexistence of attributes . . . .219 

D. Verbal and real propositions, 

63. Difficulty of saying what are ' verbal ' propositions according 

to Mill. With the formal logicians all general proposi- 
tions, which are logically justifiable, must be ' verbal ' .221 

64. With Locke, again, only singular propositions are ' instruc- 

tive ' (real), except in morals and mathematics, where 
general propositions may be so . . . . . . 221 

65. Though on his own theory he could not explain how this is 

6Q. Kant retains the doctrine that ' empirical ' judgments cannot 
be universal and necessary, (which is true, if they only 
summarise feelings) ....... 224 

67. And holds therefore that universal and necessary judgments, 

if synthetical, must be a priori ..... 224 

68. But (1) his ' analytical ' judgments (unlike Locke's ' verbal ') 

are not mere analyses of names, and (2) his ' synthetical ' 
judgments (unlike Locke's l real ') imply conceptions not 
of empirical origin . . . . . . . . 225 

69. Must we not then modify (1) Kant's distinction between 

mathematical and other knowledge, and (2) his doctrine 
that all extension of knowledge implies intuition ? . . 226 

70. ' Body is extended ' involves a synthesis as much as ' body 

has weight ' ; the difference lies Only in the complexity of 

the synthesis ......... 227 

71. I.e. while both may represent mere analyses of conceptions, 

yet (1) neither is originally arrived at without experience 

or an intellectual synthesis 228 

a 2 



72. (2) To a perfect intelligence both would be equally neces- 

sary ; and (3), as truths, the distinction is only between 

the more elementary and the more complex . . . 229 

73. In what sense can the distinction between judgments re- 

specting matters of fact and respecting ideas be maintained ? 
Locke and Hume ........ 230 

74. If, as Mill holds, there are universal propositions about 

matters of fact, these must be, not events of feeling, but 
elements in an intelligible order . . . . .231 

75. For the above distinctions of Locke and Kant it is better to 

adopt another of Kant, that between propositions which 
do and those which do not concern objects of possible ex- 
perience 231 

E. Definition. 

76. The ambiguity in Mill's view of propositions comes out in 

his account of definitions ...... 233 

77. He identifies them with ' verbal ' propositions, and yet dis- 

tinguishes ' complete ' from ' incomplete,' and requires the 
former to state ' the whole of the facts ' . . . . 233 

78. In regard to scientific definitions, which ' expound a classifi- 

cation/ he virtually adopts the view of Plato and Aristotle 235 

79. He is afraid that if he admits definition to be of the ' nature 

of things,' he implies the reality of ' essences ' . . 235 

80. Every definition explains a conception represented by a name, 

but the question is, what is the conception and in what 
sense does the definition explain it ? . . . . 236 

81. Definitions of current terms, of scientific genera, of metaphy- 

sical and of legal or moral principles, must all in different 
ways be incomplete . . 236 

F. Space and geometrical truth. 

82. The question as to the nature of mathematical definitions 

turns on the view taken of space. (What is accounted for 
by Spencer and others is not space, but our beliefs about 
space) . 238 

83. Space according to Berkeley and Hume . . . . 238 

84. The ' priority ' assigned by Kant to space and time is com- 

patible with the historical view of the origin of the ideas 

of them 239 

85. In his definition of space as ' form of outer sense, < form' = 

the relation by which the mind constitutes an external 
object 240 



86. And the point of the definition, which owing to the primari- 

ness of space is necessarily tautological, is that space 
does not belong to things apart from thought, nor to all 
objects of thought, nor to thought itself . . . .241 

87. The theory that space is an attribute of c matter ' either 

explains nothing (if matter is unknown), or explains it by 

1 thought ' (though not the thought of the individual) . .• 242 

88. Space is a particular mode of the relation of mutual limita- 

tion between objects : as the condition of all perceivable 
objects, it can be considered apart from all other condi- 
tions . . . .243 

89. In this sense it may be called i an abstraction ' (as Kant calls 

it ' an intuition '), but not as abstracted from sensitive 
experience ......... 243 

90. The points, lines, surfaces, which are the material of geo- 

metry, are various forms of the limit between objects, got 

by detachment of the non-sensuous element in perception 244 

91. The definitions of geometry are mental constructions out of 

this ideal material, and the axioms are involved in these 

acts of construction .... . . . . 245 

92. The definitions imply an a priori synthesis (Kant), which holds 

good of all possible experience (is in l pure intuition ') . 246 

93. Mill objects to saying that geometry rests on definitions, 

because he thinks definition has only to. do with names . 246 

94. The universality of the singular proof in geometry is due to 

the fact that we construct the figures out of elements im- 
plied in all possible experience ..... 247 

95. The propositions on which the proofs depend are neither 

statements of sensible events nor analyses of the meaning 

of names . . . . . . . . . 248 

96. The distinction between the ' necessity' of mathematical and 

the 'contingency' of physical truth is really one between 
complete and partial truth . . . . . .249 

97. It does not mean that mathematical truths are only hypo- 

thetically ' true, a theory which rests on the identification 

of the ' real ' with the ' felt ' 250 

G. Time. 

98. Time is the relation of the elements of a vanishing manifold 

which (like space) implies a permanent subject . . 252 

99. Time as the l form of inner sense,' (Kant). The distinction 

between 'outer ' and ' inner ' sense as = that between 
bodily and mental is (a) untenable in itself . . . . 252 



100. And (b) does not correspond to the Kantian distinction 

between ' outer ' and ' inner ' as = in space and in time . 254 

101. This distinction, as understood by Kant, is itself open to 

objection, and as properly understood is not well expressed 

by ' outer ' and ' inner ' . . . . . . . 254 

102. Time is the prima?'!/ condition of all feelings as simply at- 

tended to or thought ; space is their condition as not in 
time, as attended to together ..... 256 

H. Demonstration and necessary truth. 

103. To the questions- (a) from what experience are primary 

mathematical truths derived ? (b) how are they derived ? 
Mill answers (a) from sensation, (b) by generalisation . 258 

104. Where then (1) do we get the material for the generalisa- 

tion, if (as Mill holds) we never have sensible experience 

of these truths ? 259 

105. And (2) how, when made, do we apply it to reality? . . 261 

106. In his opposition of knowing ' by our eyes ' to knowing ' by 

intuition ' Mill is right as against Whewell, who takes 

* intuition' to = imaginary looking . . . .262 

107. It is an abstract view of space which says ' body cannot act 

where it is not ' ; in truth it determines that which is 
where it is not by not being there 263 

108. The antithesis between * necessary ' truths and truths ' of 

experience ' is misleading. All general truths, so far as 
unqualified, are necessary ...... 264 

109. ' Inconceivability of the contradictory ' is not the test of the 

truth of a proposition, because it is simply equivalent to 

its truth . 265 

110. Mill is right in denying it, but denies it on wrong grounds. 267 

111. If 'inconceivable' = 'unbelievable' in Mill's sense, no 

scientific theories are believable . . . . .268 

112. The truth is that these theories are not theories about the 

reality ; they are the reality 268 

113. Spencer's doctrine of inconceivability, (a) as applied to 

1 belief in our own sensations.' What he calls ' incon- 
ceivable ' is really the necessary complement of the con- 
ception which it negatives 269 

114: (b) As applied to belief in an objectively external world. 
He confuses 'objective' with 'external,' and subjects 
thought to conditions of its own making . . . . 270 

115. Mill's doctrine that truths of number are got by 'generali- 
sation' assumes what it has to explain, the act of 
counting . . . ... . . . 271 


T. Syllogism. 


116. In his theory of syllogism Mill is right as against the theory 

which he attacks ........ 273 

117. But his own theory is open to the same objections, and 

these cannot be answered by denying ' general truths ' 
about nature ........ 274 

118. The true answer is that inference has to do, not with the 

quantity of particulars, but with their nature ; it is from 

the phaenomenon to its conditions ..... 275 

119. Mill's view of inference in his account of syllogism is 

inconsistent with that in his account of induction . . 276 

120. The likeness and unlikeness of inference in mathematics, 

natural science, and law respectively 277 

121. Mill's ' universal type' does not really express either the 

process or the result of inference . . . . .278 

K. Induction. 

122. Misconceptions implied in Mill's various accounts of induc- 

tion ... . . . . . . . 281 

123. They go along with a misconception of the axiom of the 

uniformity of nature, which does not mean that 'the 
future will resemble the past ' . . . . . 28 1 

124. The principle of induction cannot be derived from ' simple 

enumeration ' ; it is implicit in the most elementary act 

of knowledge ......... 283 

125. Mill's account of induction expresses only an incident, not 

the essence, of it . . . . . . . . 284 

126. He holds two inconsistent views of science, as = (1) dis- 

covery of causes, (2) registration of resemblances. In- 
terest of his controversy with Whewell . . . . 285 

127. What he distinguishes as 'observation,' 'description,' and 

'induction,' are really stages in one act by which the 
world becomes known to us . . . . . . 286 

128. Observation either does not involve resemblance at all, or 

involves more than resemblance . . . . . 286 

129. The ordinary view ignores the fact that resemblance implies 

an act of colligation which itself constitutes the resem- 
bling objects 287 

130. All science is a progressive correlation, i.e. conditioning, 

of feelings 288 

131. The phenomena between which ' uniformities ' are said by 

Mill to be discovered, are themselves sums of conditions . 290 



132. Mill doe3 not draw the right distinction between ' descrip- 

tion ' (as implying ' abstraction ') and ' explanation ' (as 
implying ' generalisation ')...... 290 

133. It is true that we get our conceptions from the ' facts,' but 

the facts from which we get them are our own imperfect 
observations . . . . . . . . . 292 

134. Thus neither was Kepler's discovery an * abstraction ' in 

the ordinary S( ns3, nor that of Newton a l generalisation ' 

in the ordinary sense . . . . . . .293 

135. The real difference between them lies in the relative com- 

pleteness of the colligations . . . . . . 294 

L. Causation. 

136. Ordinary views as to the alternative theories of causation . 296 

137. Hume's doctrine, and its true logical outcome . . . 297 

138. Scientific men, who suppose themselves to adopt it, really 

adopt only a residuum of it . . . . . . 299 

139. Viz. that no reason can be given why one event should 

follow another. Supposed antithesis between ' reason ' 

and < cause ' 300 

140. It is true that no fact can be proved by reasoning, but 

equally true that nothing is a fact which does not imply 
reasoning . . . . . . . . . 301 

141. If ' cause ' = ' sum of conditions,' it cannot = 'antecedent 

event' 302 

142. It may be said that the sum of conditions is resolvable into 

certain sequences and simultaneities .... 303 

143. But the relation of a phenomenon to all sequences and 

simultaneities, which alone is its adequate cause, implies 
something not successive or simultaneous . . . . 304 

144. Mill inconsistently holds that the relation of cause and 

effect is an objective relation, and yet that it consists in 
sequence of feelings . . . . . .305 

On the different senses of ' FREEDOM ' AS applied 


1. In one sense (as being search for self- satisfaction) all will is 

free ; in another (as the satisfaction sought is or is not 
real) it may or may not be free . . . . . . 308 

2. As applied to the inner life ' freedom ' always implies a 

metaphor. Senses of this metaphor in Plato, the Stoics, 

* St - P*»* . . . . 309 



3. St. Paul and Kant. It would seem that with Kant ' free- 

dom ' means merely consciousness of the possibility of it 

(' knowledge of sin ') 311 

4. Hegel's conception of freedom as objectively realised in the 

state .312 

5. It is true in so far as society does supply to the individual 

concrete interests which tend to satisfy the desire for per- 
fection . . . . . . . . . . 312 

6. Though (like the corresponding conception in St. Paul) it is 

not and could not be realised in any actual human society 314 

7. In all these uses 'freedom' means, not mere self-determi- 

nation or acting on preference, but a particular kind of 
this 315 

8. The extension of the term from the outer to the inner rela- 

tions of life, though a natural result of reflection, is apt to 

be misleading ........ 315 

9. Thus the question, Is a man free ? which may be properly 

asked in regard to his action, cannot be asked in the same 
sense in regard to his will ... . . . . . 316 

10. The failure to see this has led to the errors (1) of regarding 

motive as something apart from and acting on will, (2) of 
regarding will as independent of motive . . . .317 

11. Thus the fact that a man, being what he is, must act in a 

certain way, is construed into the negation of freedom . . 318 

12. And to escape this negation recourse is had to the notion of 

an unmotived will, which is really no will at all . .319 

13. The truth is that the will is the man, and that the will 

cannot be rightly spoken of as ' acting on ' its objects or 
vice versa, because they are neither anything without the 
other 319 

14. If however the question be persisted in, Has a man power 

over his will ? the answer must be both ' yes ' and ' no ' . 320 

15. ' Freedom ' has been taken above (as by English psycholo- 

gists generally) as applying to will, whatever the character 

of the object willed 320 

16. If taken (as by the Stoics, St. Paul, Kant (generally), and 

Hegel) as applying only to good will, it must still be recog- 
nised that this particular sense implies the generic . .321 

17. Whatever the propriety of the term in the particular sense, 

both ' juristic ' and ' spiritual ' freedom spring from the 
same self- asserting principle in man 322 

18. And though the former is only the beginning of full freedom, 

this identity of source will always justify the use of the 
word in the latter sense ....... 323 



19. But does not the conception of ' freedom ' as = the moral 

ideal imply an untenable distinction like that of Kant be- 
tween the ' pure ' and ' empirical * ego ? ... 324 

20. The 'pure' and 'empirical' ego are one ego, regarded (1) 

in its possibility, (2) as at any given time it actually is . 325 

21. In man the self-realising principle is never realised ; i.e. the 

objects of reason and will only tend to coincide . . 326 

22. So far as they do coincide, man may be said to be' free ' and 

his will to be ' autonomous ' 327 

23. The growing organisation of human life provides a medium 

for the embodiment, and disciplines the natural impulses 

for the reception, of the idea of perfection . . • • 328 

24. The reconciliation of reason and will takes place as the indi- 

vidual more and more finds his own self-satisfaction in 
meeting the requirements of established morality . . 330 

25. Until these come to be entirely superseded by the desire of 

perfection for its own sake, and his will becomes really free 331 

Lectures on the Principles op Political Obligation. 

A. The grounds of political obligation. 

1. Subject of the inquiry ........ 335 

2. Its connection with the general theory of morals. Ideal 

goodness is to do good for its own sake : but there must 
be acts considered good on other grounds before they can 
be done for the sake of their goodness .... 335 

3. When however the ideal comes to be recognised as the ideal, 

the lower interests and rules must be criticised and revised 

by it 336 

4. The criticism of interests will yield a ' theory of moral sen- 

timents '; that of rules will relate (1). to positive law, (2) 

to the law of opinion ....... 337 

5. As moral interests greatly depend on recognised rules of 

conduct, and these again on positive law, it is best to 
begin by considering the moral value of existing civil in- 
stitutions .......... 337 

6. The condition of morality is the possession of will and reason, 

and it is realised in a personal character in which they are 
harmonised 337 

7. Civil institutions are valuable so far as they enable will and 

reason to be exercised, and so far they answer to ' jus 
naturae ' 338 



8. The essential questions as to the ' law of nature ' are, (1) are 

there rights and obligations other than those actually en- 
forced ? (2) If so, what is the criterion of them ? . . 389 

9. While rejecting the theory of a ' state of nature,' we may 

still use ' natural' of those rights which ought to be, though 
they actually are not ........ 339 

10. Such 'natural law' is (as admitting enforcement) distinct 

from, but (as implying a duty to obey it) relative to, the 
moral law . . . . . . . . . 340 

11. Hence two principles for the criticism of law, (1) only ex- 

ternal acts can be matter of obligation proper, (2) the ideal 
of law must be determined by reference to the moral end 
which it serves ......... 340 

12. Observe (a) that in confining law to ' external actions,' we 

mean by 'actions' intentions, without which there is properly 

no ' action '......... 341 

13. (b) That by ' external ' we mean that law, though it does 

supply motives to action, looks merely to whether the 
action is done, not to whether it is done from a particular 
motive 342 

14. Law then can only enjoin or forbid certain acts; it cannot 

enjoin or forbid motives ....... 343 

15. And the only acts which it ought to enjoin or forbid are those 

of which the doing or not doing, from whatever motive, is 
necessary to the moral end of society ..... 343 

16. The principle of 'natural law,' then, should be to enjoin all 

acts which further action from the highest motive, and no 

acts which interfere with such action .... 344 

17. This principle would condemn much legislation which has 

tended, e.g., to weaken religion, self-respect, or family 
feeling 344 

18. This, and not the principle of 'laissez-faire,' is the true 

ground of objection to ' paternal government ' . . . 345 

19. The theory of political obligation (i.e. of what law ought to 

be, and why it ought to be obeyed) is not a theory (a) as 

to how existing law has come to be what it is . . . 346 

20. Nor (b) as to how far it expresses or is derived from certain 

original ' natural ' rights . . . . . . .346 

21. ' Natural' rights (like law itself) are relative to moral ends, 

i.e. they are those which are necessary to the fulfilment of 
man's moral vocation as man . . . . . . 347 

22. This however is not the sense in which political obligation 

was based on ' natural rights ' in the seventeenth and 
eighteenth centuries, previously to utilitarianism . .347 


23. The utilitarian theory so far agrees with that here advocated 

that it grounds existing law, not on a * natural ' law prior 

to it, but on an end which it serves . . . . . 348 

24. The derivation of actual rights from natural (i.e. more primi- 

tive) rights does not touch the real question, viz. how there 
came to be rights at all 349 

25. The conception of a moral ideal (however dim) is the con- 

dition, of the existence of rights, and conversely anyone 

who is capable of such a conception is capable of rights . 350 

26. Thus the consciousness of having rights is co-ordinate with 

the recognition of others as having them, the ground of both 
being the conception of a common good which ought to be 
attained. ......... 351 

27. Eights then can only subsist among ' persons,' in the moral 

sense of 'persons,' i.e. beings possessed of rational will . . 351 

28. Though the moral idea of personality is later in formulation 

than the legal, and this again than the actual existence of 
rights 352 

29. Rights which are directly necessary to a man's acting as a 

moral person at all may be called in a special sense * per- 
sonal' 353 

30. Nor is there any objection to calling them ' innate ' or ' na- 

tural,' if this means * necessary to the moral development 

of man,' in which sense ' duties ' are equally ' natural ' . 353 

31. Without a society conscious of a common interest there can 

be only ' powers,' no ' rights ' 354 

B. Spinoza. 

32. Spinoza, seeing that 'jus naturae ' = ' potentia,' and not see- 

ing that it is not really 'jus' at all, identifies all 'jus' with 

' potentia,' both in the state and in the individual . . 355 

33. From which it follows that the ' right ' of the state against its 

individual members is only limited by its 'power' . . 357 

34. And the same principle applies to the relations of one state 

to other states ........ 357 

35. But, according to Spinoza, though everything is ' lawful' for 

the state, everything is not ' best,' and the ' best ' state is 
that which secures a life of ' peace,' i.e. rational virtue or 
perfection .......... 35B 

36. This conclusion does not seem consistent with his starting- 

point, according to which men are ' naturally enemies ' . 359 

37. From such a 'status naturalis' there is no possible transition 

to the ' status civilis,' and the phrase l jus natural ' remains 
unmeaning , . . , . , , . . 361 



38. Spinoza's error of regarding l rights ' as possible apart from 

society was confirmed by his denial of final causes . .362 

39. It was just because Plato and Aristotle regarded man as find- 

ing his end in the end of the state, that they founded a true 
. theory of rights . . . . . . . . . 363 

40. Spinoza, however, while insisting that man is ' part of nature,' 

yet places his 'good' in understanding nature and so 
acquiring a new character . . . . . .363 

41. In thus recognising the idea of perfection as a determinant 

of life, he really recognises an operative final cause, though 
without seeing its bearing on the theory of right . . . 365 

C. Hobbes. 

42. Hobbes differs from Spinoza in regarding the right of the 

sovereign, not as limited by his power, but as absolute . 366 

43. Statement of his doctrine ....... 367 

44. He uses ' person,' as in Roman law, for either (1) a complex 

of rights, or (2) the subject of those rights . . . 367 

45. Though by his theory the sovereign may be one or many, 

and sovereignty is transferable by the act of a majority, he 
tacitly vindicates the absolute right of a de facto monarchy 368 

46. The radical fiction in his theory is that there can be any 

1 right ' after the institution of sovereignty, if (as he holds) 
there is none before it ....... 369 

47. To justify his doctrine of absolute submission he has to 

assume a ' law of nature ' which binds men to keep cove- 
nant, while yet he holds the ' law of nature ' to be mere 
1 power ' and covenants to be only valid under an imperium 370 

48. His ' contract ' can confer none but natural right, and that is 

either not a right at all, or (if it is) it belongs to all men, 
subject and sovereign alike . . . . . .371 

49. The real flaw in the theory of contract is not that it isunhis- 

torical, but that it implies the possibility of rights and 
obligations independently of society ..... 372 

50. Though it has not been popularly accepted as regards the 

rights of sovereigns over subjects, the behaviour of indivi- 
duals to society is to a great extent practically determined 
by it 373 

D. Locke. 

51. The development of this latter side of it is peculiarly due to 

Rousseau, but Locke, Hooker, and Grotius have essentially 

the same conception : Spinoza alone differs . . . . 374 



52. Ambiguity of their phrase * state of nature.' They agree in 

treating it as the negation of the ' political state.' But if 

so, contract would be impossible in it . . . . . 375 

53. Nor could it be a state of ' freedom and equality,' as most of 

them assume it to be 375 

54. And if this state of nature implies consciousness of obliga- 

tion, it must imply recognition of social claims, and must 
therefore be virtually a political state 376 

55. In fact the theory of a state of nature governed by a law of 

nature, as preceding civil society, must be untrue either to 

the conception of law or to that of nature .... 377 

56. Locke differs from Hobbes (1) in distinguishing the ' state of 

nature ' from the ' state of war ' 378 

57. He implies (more consistently than Hobbes) that the ' state 

of nature ' is one in which the ' law of nature ' is observed 379 

58. (2) He limits the supreme power in the state by the legis- 

lature, which holds its functions in trust from the com- 
munity .......... 380 

59. And this distinction between the supreme community and 

the supreme executive enables him to distinguish between 
dissolution of the political society and dissolution of the 
government, which Hobbes had confused . . .381 

60. He invests the community with the right of assuming the 

powers which they have delegated, and thus justifies revo- 
lution when it is the act of the whole community . . . 381 

61. The difficulty is to determine when it is the act of the whole 

community, and on this Locke's theory gives no help . 382 

62. The difficulty indeed is not so great as that of conceiving 

the act of original devolution of power, and is inherent in 

the theory of contract ....... 383 

63. In the particular case of the reform of the English represen- 

tative system, Locke does not contemplate the carrying 

out of his own theory . . . - . . . . 384 

E. Rousseau. 

64. Rousseau conceives the community to be in continual 

exercise of the power which Locke conceives it to have 
exercised once and to hold in reserve . .... 386 

65. In his view of the motive for passing from the state of 

nature into the civil state he is more like Spinoza than 
Locke 386 

66. His statement of the origin and nature of the ' social con- 

tract' 386 



67. Its effects upon the individual . . . . . . 387 

68. His idea of the sovereign is really that of a supreme disin- 

terested reason, but he fuses this with the ordinary idea of 

a supreme coercive power ...... 388 

69. The practical result of his theory has been a vague exalta- 

tion of the will of the people, regardless of what ' the 
people ' ought to mean 388 

70. Further consequences of his ideal conception of sovereignty. 

It cannot be alienated, represented, or divided . . . 389 

71. Thus the ' government' is never the same as the ' sovereign,' 

and constitutions differ according to where the government, 

not the sovereignty, resides . . . . . . 390 

72. The institution of government is not by contract, but by the 

act of the sovereign, and this act must be confirmed or 
repealed periodically ...... 391 

73. His distinction between the ' will of all' and the 'general 

will ' : the latter always wills the common good, though it 
may be mistaken as to means 392 

74. He admits however that it may be overpowered by particular 

interests, and so find no expression even in the vote of a 
general assembly ........ 393 

75. What then is the test of the ' general ' will ? Absolute 

unanimity is what Kousseau requires of the parties to the 
original contract ........ 394 

76. But what is to decide whether their successors are parties 

to it ? Not ' residence,' unless there is also freedom to 
migrate 395 

77. The element of permanent value in Rousseau is his concep- 

tion of the state as representing the ' general will ' . . 395 

78. Difficulties in this conception. It seems that either no actual 

state realises it, or that there may be a state without a 

true sovereign . . . . . . . . . 396 

79. We may distinguish between de facto and de jure sove- 

reignty, and say that Rousseau meant the latter ; but this 

is only an inference from what he says .... 396 

F. Sovereignty and the general will. 

80. Hence it may be asked, (1) Is any actual sovereignty 

founded on the \ general will ' ? (2) Can sovereignty de 
jure be truly said to be founded on it ? (3) If so, must it 
be expressed through the vote of a sovereign people ? . . 399 

81. (1) According to (e.g.) Austin's definition of sovereignty, 

we should answer this question in the negative . . 399 



82. (Observe that from Austin's definition it would follow that, 

while every ' law' implies a « sovereign,' a ' sovereign's ' 
commands need not be ' laws ') . . . ... 400 

83. That definition directly contradicts that of Rousseau, in (a) 

placing sovereignty in determinate persons, (b) making its 
essence lie in power to compel obedience .... 401 

84. Actual sovereignty combines both definitions; the habitual 

obedience of subjects to the sovereign is due to the 
sense that by obeying they secure certain ends . . . 402 

85. So far as Austin means that a fully developed state implies a 

determinate supreme source of law, he is right as against 
Rousseau . . 403 

86. But if sovereign power = the aggregate influences which 

really make the people obedient, it must be sought in the 

1 general will ' . . 404 

87. Such power need not be ' sovereign ' in the narrower sense, 

and may coexist with a separate coercive power which is 

' sovereign ' 404 

88. This was the case in ancient despotisms, and is so in the 

modern empires of the East 405 

89. So in states under foreign dominion, which retain a national 

life, the technical sovereign is not the law-making and law- 
maintaining power 406 

90. Under the Roman Empire, in British India,' in Russia, where 

the technical is also the real sovereign, its strength rests in 
different degrees on the general will 406 

91. Thus the answer to question (1) depends on the sense of 

1 sovereign.' If it = a power which guarantees equal 
rights, it is implied in every ' political ' society . . . 408 

92. But (a) it need not be the supreme coercive power, and (b) 

if it is so, it is not because it is so that it commands 
habitual obedience ........ 408 

93. Thus (retaining the technical use of ' sovereign ') it is true 

that if the sovereign is to be so really, it must express and 
maintain a general will ....... 409 

94. Though this is compatible with the fact that some of the laws 

of the sovereign conflict with the general will . . . 410 

95. Thus as to question (2), (above, sec. 80), if sovereignty is 

said to rest on the general will ' de jure,' either ' sovereign ' 

or 'jus' is not used in the strict sense . . . .411 

96. An antithesis between sovereign 'de jure' and ' de facto' 

can only arise from a confusion between ' sovereign ' as = 

the source of law and * sovereign' as = the 'general will' 412 



97. Though there are cases in which (in a different sense) a 

sovereign may be conveniently described as ' de facto,' not 

' de jure,' or vice versa ....... 414 

98. Similarly, to say that the people is c sovereign de jure ' is to 

confuse the general will with the coercive power of the 
majority 414 

99. Rousseau's confusion is due to the theory of { natural rights,' 

(that the individual is not bound by anything which he 

has not individually approved) . . . . .415 

100. The individual must indeed judge for himself whether a law 

is for the common good ; but though he judge it not to be, 

he ought as a rule to obey it . . . . . . 416 

101. Cases in which a doubt may arise 417 

102 (a) Where the legal authority of the law is doubtful, owing 

to the doubt where the sovereignty in the state resides . 417 

103. In such cases the truth generally is that the ' right,' on the 

particular issue, has not yet formed itself . . .418 

104. But it does not follow that because the ' right ' is on both 

sides, one is not ' better ' than the other ; though this may 

be the case 419 

105. In such cases of disputed sovereignty the distinction of 

'de jure' and ' de facto' may be applied, though it is 
better to say that the sovereignty is in abeyance . . 420 

106. The individual, having no ' right ' to guide him, should take 

the side whose success seems likely to be best for man- 
kind 421 

107. (b) Another case is where there is no legal way of getting 

a bad law repealed. Here it is a question, not of right, 

but of duty , to resist the sovereign. .... 422 

108. Nor is it a question of the right of a majority, as a majority, 

to resist : it may be the duty of a helpless minority . . 422 

109. Some general questions which the good citizen may put to 

himself in such dilemmas ...... 424 

110. They can, indeed, seldom be applied by the agents at the 

time as they can be after the event . .... 424 

111. In simple cases we may judge of the right or wrong of an 

act by the character which it expresses, but generally we 

can only judge them by its results .... 424 

112. All that the historian can say is that on the whole the best 

character is likely to produce the best results, notwith- 
standing various appearances to the contrary . . . 425 



G. Will, not force, is the basis of the state. 


113. The doctrines which explain political obligation by con- 

tract agree in treating sovereign and subject apart, 
whereas they are correlative . . . . . . 427 

114. For the desire for freedom in the individual is no real desire 

unless he is one of a society which recognises it. (Slaves 

are not a real exception to this) . . . . . 428 

115. And without an authority embodied in civil institutions he 

would not have the elementary idea of right which enables 

him to question the authority 429 

116. But the theory of contract expresses, in a confused way, the 

truth that only through the common recognition of a 
common good, and its embodiment in institutions, is 
morality possible . . . . • . . .429 

117. Thus morality and political subjection have a common 

source .......... 430 

118. And both imply the twofold conception, (a) l I must though 

I do not like,' (b) ' I must because it is for the common 
good which is also my good ' 431 

119. It is a farther and difficult question, how far the sense of 

common interest can be kept alive either in the govern- 
ment or subjects, unless the people participates directly in 
legislation ......... 432 

120. And this suggests the objection, Is it not trifling with words 

to speak of political subjection in modern states as based 

on the will of the subjects ? . . . . . . 433 

121. We must admit (a) that the idea of the state as serving a 

common interest is only partially realised, even by the 
most enlightened subject, though so far as realised it is 
what makes him a loyal subject . .... 434 

122. (b) That if he is to be an intelligent patriot as well as a 

loyal subject, he must take a personal part in the work of 

the state 435 

123. And (c) that even then his patriotism will not be a passion 

unless it includes a feeling for the state analogous to that 
which he has for his family and home . . . . 43 G 

124. But are we not again assuming what was disputed, viz. that 

a sense of its serving a common interest is necessary to 

the existence of the state ?...... 437 

125. Observe that the idea of an end or function, realised by 

agencies unconscious of it and into which it cannot be re- 
solved, is already implied even if the state be treated as a 
' natural organism ' , . . 437 



126. Such a treatment, however, would ignore the distinction 

between the ' natural ' and the ' human ' or ' moral ' agencies 
which have operated in the production of states . .438 

127. It may be objected that these l human' agencies are not 

necessarily * moral,' but on the contrary are often selfish . 439 

128. But though human motives are never unalloyed, they only 

produce good results so far as they are fused with and 
guided by some unselfish element . . . . .439 

129. If e.g. we would form a complete estimate of Napoleon, we 

must consider not only his ambition but the particular 
form in which his ambition worked ... . . 440 

130. And further reflect that the idiosyncrasy of such men plays 

but a small part in the result, which is mainly due to agencies 

of which they are only the most conspicuous instruments 441 

131. Thus an ideal motive may co-operate with the motives of 

selfish men, and only through such co-operation are they 
instrumental for good . . . . . . . 441 

132. The fact that the state implies a supreme coercive power 

gives colour to the view that it is based on coercion ; 
whereas the coercive power is only supreme because it is 
exercised in a state, i.e. according to some system of law, 
written or customary . . . . . . . 442 

133. In the absence of any other name, ' state ' is the best for a 

society in which there is such a system of law and a power 

to enforce it . . . . . . . . 444 

134. A state, then, is not an aggregate of individuals under a 

sovereign, but a society in which the rights of men already 
associated in families and tribes are defined and har- 
monised ......... 445 

135. It developes as the absorption of fresh societies or the ex- 

tended intercourse between its members widens the range 

of common interests and rights ..... 445 

136. The point to be insisted on is that force has only formed 

states so far as it has operated in and through a pre-exist- 
ing medium of political, tribal, or family ' rights ' . . 446 

H. Has the citizen rights against the state ? 

137. As long as power of compulsion is made the essence of the 

state, political obligation cannot be explained either by 
the theory of * consent,' or by that which derives all right 
from the sovereign ....... 448 

138. The state presupposes rights, rights which may be said to 

belong to the ' individual ' if this mean * one of a society 

of individuals * . . . . . . . . • 449 



139. A right may be analysed into a claim of the individual upon 

society and a power conceded to him by society, but 
really the claim, and the concession are sides of one and 
the same common consciousness ..... 450 

140. Such common consciousness of interests is the ground of the 

1 natural right ' of slaves and of the members of other states 450 

141. But though in this way there may be rights outside the 

state, the members of a state derive the rights which they 
have as members of other associations from the state, and 
have no rights against it . . . . . .451 

142. I.e. as they derive their rights from their membership in 

the state, they have no right to disobey the law unless it 

be for the interest of the state . . . ... 452 

143. And even then only if the law violates some interest which 

is implicitly acknowledged by the conscience of the com- 
munity .... ..... 454 

144. It is a further question when the attempt to get a law re- 

pealed should be exchanged for active resistance to it . . 455 

145. E.g. should a slave be befriended against the law ? The slave 

has as a man certain rights which the state cannot extin- 
guish, and by denying which it forfeits its elaim upon him . 457 

146. And it may be held that the claim of the slave upon the 

citizen, as a man, overrides the claim of the state upon 
him, as a citizen ........ 458 

147. Even here, however, the law ought to be obeyed, supposing 

that its violation tended to bring about general anarchy . 459 

I. Private rights. The right to life and liberty, 

148. There are rights which men have as members of associa- 

tions, which come to be comprised in the state, but which 

also exist independently of it . . . . .460 

149. These are ' private ' rights, divided by Stephen into (a) 

personal, (b) rights of property, (c) rights in private 
relations .......... 460 

150. All rights are ' personal ' ; but as a man's body is the con- 

dition of his exercising rights at all, the rights of it may 

be called * personal' in a special sense .... 461 

151. The right of * life and liberty ' (better, of ' free life '), being 

based on capacity for society, belongs in principle to man 

as man, though this is only gradually recognised . . 461 

152. At first it belongs to man as against other members of his 

family or tribe, then as against other tribes, then as 
against other citizens, which in antiquity still implies 
great limitations . 462 



153. Influences which have helped to break down these limita- 

tions are (a) Roman equity, (b) Stoicism, (c) the Christian 
idea of a universal brotherhood . . . . . 463 

154. This last is the logical complement of the idea that man as 

such has a right to life ; but the right is only negatively 
recognised in modern Christendom .... 463 

155. It is ignored e.g. in war, nor is much done to enable men to 

fulfil their capacities as members of humanity . . . 464 

156. Four questions as to the relation of the state to the right of 

man as man to free life . . . . . . ,466 

K. The right of the state over the individual in war, 

157. (1) Has the state a right to override this right in war ? It 

must be admitted that war is not ' murder,' either on the 
part of those who fight or of those who cause the war . . 466 

158. Yet it may be a violation of the right of life. It does not 

prove it not to be so, that (a) those who kill do not intend 

to kill anyone in particular ...... 467 

159. Or that (b) those who are killed have incurred the risk 

voluntarily. Even if they have, it does not follow that 
they had a ' right ' to do so . . . . . . 468 

160. It may be said that the right to physical life may be over- 

ridden by a right arising from the exigencies of moral life 470 

161. But this only shifts the blame of war to those who are re- 

sponsible for those exigencies ; it remains a wrong all the 
same 470 

162. But in truth most wars of the last 400 years have not been 

wars for political liberty, but have arisen from dynastic 
ambition or national vanity . . . . , .471 

163. Admitting, then, that virtue may be called out by war 

and that it may be a factor in human progress, the 
destruction of life in it is always a wrong . . . 473 

164. * But if it be admitted that war may do good, may not those 

who originate it have the credit of this ? ' . . . . 474 

165. If they really acted from desire to do good, their share in 

the wrong is less ; but in any case the fact that war was 
the only means to the good was due to human agency, 
and was a wrong ........ 474 

166. (2) (See sec. 157). Hence it follows that the state, so far 

as it is true to its principle, cannot have to infringe the 
rights of man as man by conflicts with other states . . 476 

167. It is not because states exist, but because they do not fulfil 

their functions as states in maintaining and harmonising 
general rights, that such conflicts are necessary . . 477 



168. This is equally true of conflicts arising from what are called 

* religious ' grounds ........ 478 

169. Thus no state, as such, is a bsolutely justified in doing a 

■wrong to mankind, though a particular state may be 
conditionally justified 479 

170. It may be objected that such a * cosmopolitan ' view ignores 

the individuality of states, and could only be realised if 
they were all absorbed in a universal empire . . . 480 

171. It is true that public spirit, to be real, must be national; 

but the more a nation becomes a true state, the more 
does it find outlets for its national spirit other than con- 
flicts with other nations 481 

172. In fact the identification of patriotism with military aggres- 

siveness is a survival from a time when states in the full 
sense did not exist ........ 482 

173. And our great standing armies are due, not to the develop- 

ment of a system of states, but to circumstances which 
witness to the shortcomings of that system . . . 482 

174. The better the organisation of each state, the greater is the 

freedom of communication with others, especially in 
trade, which, beginning in self-interest, may lead to the 
consciousness of a higher bond . . . . . . 483 

175. As compared with individuals, any bonds between nations 

must be weak ; on the other hand, governments have less 
temptation than individuals to deal unfairly with one 
another 484 

L. The right of the state to punish. 

176. (3) (See sec. 157). What right has the state to punish? 

The right to live in a community rests on the capacity to 
act for the common good, and implies the right to protect 
such action from interference . . - . . . . 486 

177. A detailed theory of punishment implies a detailed theory 

of rights. Here we can only deal with principles . . 486 

178. Is punishment retributive ? Not in the sense that it carries 

on a supposed * right ' of private vengeance, for no such 

1 right ' can exist 487 

179. The most rudimentary ' right ' of vengeance implies social 

recognition and regulation, in early times by the family . 488 
] 80. And its development up to the stage at which the state 
alone punishes is the development of a principle implied 
from the first ......... 488 

181. ' But if punishment excludes private vengeance, how can it 



be retributory at all ? And how can a wrong to society 

be requited V 489 

182. When a wrong is said to be ' done to society,' it does not 

mean that a feeling of vindictiveness is excited in the 
society .......... 489 

183. The popular indignation against a great criminal is an 

expression, not of individual desire for vengeance, but of 

the demand that the criminal should have his due . . 490 

184. And this does not mean an equivalent amount of suffering ; 

nor such suffering as has been found by experience to 
deter men from the crime . . . . . . . 491 

185. Punishment, to be just, implies (a) that the person punished 

can understand what right means, and (b) that it is some 
understood right that he has violated .... 492 

186. He will then recognise that the punishment is his own act 

returning on himself; (it is in a different sense that the 
physical consequences of immorality are spoken of as a 
'punishment') 492 

187. Punishment may be said to be preventive, if it be remem- 

bered (a) that what it * prevents ' must be the violation 
of a real right, and (b) that the means by which it 'pre- 
vents ' must be really necessary . . . . . 494 

188. Does our criterion of the justice of punishment give any 

practical help in apportioning it ? . . . . 495 

189. The justice of punishment depends on the justice of the 

system of rights which it is to maintain .... 495 

190. The idea that ' just' punishment is that which = the crime 

in amount confuses retribution for the wrong to society 
with compensation for damages to the individual . . 496 

191. ' But why not hold that the pain of the punishment ought 

to = the moral guilt of the crime ? ' . . . . 497 

192. Because the state cannot gauge either the one or the other; 

and if it could, it would have to punish every case differ- 
ently 497 

193. In truth the state has regard in punishing, not primarily to 

the individuals concerned, but to the future prevention of 
the crime by associating terror with it in the general 
imagination ......... 497 

194. The account taken of ' extenuating circumstances ' may be 

similarly explained ; i.e. the act done under them requires 
little terror to prevent it from becoming general . . 498 

195. ' But why avoid the simpler explanation, that extenuating 

circumstances are held to diminish the moral guilt of the 
act?' 500 



196. Because (a) the state cannot ascertain the degree of moral 

guilt involved in a crime ; (b) if it tries to punish immo- 
rality (proper), it will check disinterested moral effort . 500 

197. Punishment, however, may be truly held to express the 

* moral disapprobation ' of society, but it is to the external 

side of action that the disapprobation is directed . . 501 

198. The principle that punishment should be regulated by the 

importance of the right violated explains the severity with 
which ' culpable negligence ' is punished . . . . 503 

199. And the punishment of crimes done in drunkenness illus- 

trates the same principle . . . . . .503 

200. It also justifies the distinction between ' criminal ' and 

1 civil ' injuries, (which is not a distinction between 
injuries to individuals and to the community, for no 

* right ' is violated by injury done to an individual as 
such) . 504 

201. There would be no reason in associating terror with breaches 

of a right which the offender either did not know that he 

was breaking or which he could not help breaking . . 505 

202. When such ignorance and inability are culpable, it depends 

on the seriousness of the wrong or the degree to which 
the civil suit involves deterrent effects, whether they 
should be treated as crimes . . . . . . 506 

203. Historically, the state has interfered first through the civil 

process ; gradually, as public alarm gets excited, more and 
more offences come to be treated as crimes . . . 507 

204. Punishment must also be reformatory, (this being one way 

of being preventive), i.e. it must regard the rights of the 
criminal ......... 508 

205. Capital punishment is justifiable only (a) if it can be shown 

to be necessary to the maintenance of society, (b) if there 
is reason to suppose the criminal to be permanently inca- 
pable of rights . . . . ... .... 509 

206. Punishment, though directly it aims at the maintenance of 

rights, has indirectly a moral end, because rights are con- 
ditions of moral well-being 510 

M. The right of the state to promote morality. 

207. (4) (See sec. 156). The right of free life is coming to be 

more and more recognised amongst us negatively \ is it 
reasonable to do so little positively to make its exercise 
possible? ......... 512 



208. First observe that the capacity for free life is a moral ca- 

pacity, i.e. a capacity for being influenced by a sense of 
common interest ; this influence will only be weakened by 
substituting for it that of law . . . . . . 512 

209. Still the state can do more than it usually does without 

deadening spontaneous action ; e.g. ' compulsory education ' 
need not be ' compulsory ' except to those who have no 
spontaneity to be deadened . . . . . .514 

210. So too with interference with ' freedom of contract ' ; we 

must consider not only those who are interfered with, but 
those whose freedom is increased by the interference . . 515 

N. The right of the state in regard to property. 

211. As to property two questions have to be kept distinct, (a) 

how there has come to be property, (b) how there has come 
to be a right of property. Each of these again may be 
treated either historically or metaphysically . •■'■■■•". 517 

212. The confusion of these questions and methods has given 

rise either to truisms or to irrelevant researches as to the 
nature of property 518 

213. Property implies (a) appropriation, i.e. an act of will, of a 

permanent self demanding satisfaction and expression . , 518 

214. (b) Kecognitition of the appropriation by others. This 

recognition cannot be derived from contract (Grotius), or 
from a supreme force (Hobbes) . . . . .519 

215. Locke rightly bases the right of property on the same ground 

as the right to one's own person ; but he does not ask what 

that ground is 522 

216. The ground is the same as that of the right of life, of which 

property is the instrument, viz. the consciousness of a 
common interest to which each man recognises every other 
man as contributing 522 

217. Thus the act of appropriation and the recognition of it con- 

stitute one act of will, as that in which man seeks a good 

at once common and personal . . . . . . 523 

218. The condition of the family or clan, in which e.g. land is 

held in common, is not the negation, but on the contrary 

the earliest expression of the right of property . . . 523 

219. Its defect lies (a) in the limited scope for free moral de- 

velopment which it allows the associates, (b) in the limited •£• 
range of moral relations into which it brings them . . 524 

220. But the expansion of the clan into the state has not brought 

with it a corresponding emancipation of the individual. 
Is then the existence of a practically propertyless class in 
modern states a necessity, or an abuse.? .... 525 



221. In theory, everyone who is capable of living for a common 

good (whether he actually does so or not) ought to have 

the means for so doing : these means are property . . 526 

222. But does not this theory of property imply freedom of 

appropriation and disposition, and yet is it not just this 
freedom which leads to the existence of a propertyless 
proletariate ? . . . . . . . .526 

223. Property, whether regarded as the appropriation of nature 

by men of different powers, or as the means required for 

the fulfilment of different social functions, must be unequal 527 

224. Freedom of trade, another source of inequality, follows 

necessarily from the same view of property : fieedom of 
bequest is more open to doubt . . . . . . 528 

225. It seems to follow from the general right of a man to pro- 

vide for his future, and (with certain exceptions) to be 
likely to secure the best distribution ; but it does not imply 
the right of entail . ....... 529 

226. Returning to the question raised in sec. 220, observe (a) that 

accumulation by one man does not itself naturally imply 
deprivation of other men, but rather the contrary . . 530 

227. Nor is the prevalence of great capitals and hired labour in 

itself the cause of the bad condition of so many of the 
working classes . . . . . . . .531 

228. The cause is to be found, not in the right of property and 

accumulation, but (partly at least) in the fact that the land 

has been originally appropriated by conquest . . .531 

229. Hence (a) the present proletariate inherit the traditions of 

serfdom, and (b) under landowning governments land has 
been appropriated unjustifiably, i.e. in various ways pre- 
judicial to the common interest . . ... 532 

230. And further the masses crowded through these causes into 

large towns have till lately had little done to improve 
their condition . . . . . . . 533 

231. Whether, if the state did its duty, it would still be advisable 

to limit bequest of land, is a question which must be 
differently answered according to circumstances . . . 534 

232. The objection to the appropriation by the state of ' un- 

earned increment' is that it is so hard to distinguish 
between ' earned ' and ' unearned ' 535 

O. The right of the state in regard to the family. 

233. The rights of husband over wife and father over children 

are (a) like that of property in being rights against all 



the world, (b) unlike it in being rights over persons, and 
therefore reciprocal . . . . . . 536 

234. The latter characteristic would be expressed by German 

writers by saying that both the ' subject ' and the ' object ' 

of these rights are persons ...... 537 

235. Three questions about them : (1) What makes man capable 

of family life? (2) How does it come to have rights? 

(3) What ought the form of those rights to be ? . . . 538 

236. (1) The family implies the same effort after permanent 

self-satisfaction as property, together with a permanent 
interest in a particular woman and her children . . 539 

237. The capacity for this interest is essential to anything which 

can be rightly called family life, whatever lower forms of 

life may historically have preceded it .... 539 

238. (2) The rights of family life arise from the mutual recogni- 

tion of this interest by members of the same clan (in 
which the historical family always appears as an element) 540 

239. Its development has been in the direction (a) of giving all 

men and women the right to marry, (b) of recognising the 
claims of husband and wife to be reciprocal. Both these 
imply monogamy . . . . . . . . 541 

240. Polygamy excludes many men from marriage and makes 

the wife practically not a wife, while it .also prevents real 
reciprocity of rights both between husband and wife and 
between parents and children ..... 541 

241. The abolition of slavery is another essential to the develop- 

ment of the true family life, in both the above respects . 542 

242. (3) Thus the right (as distinct from the morality) of family 

life requires (a) monogamy, (b) duration through life, (c) 
terminability on the infidelity of husband or wife . . 543 

243. Why then should not adultery be treated as a crime? Be- 

cause (unlike other violations of right) it is generally in 
the public interest that it should be condoned if the 
injured person is willing to condone it . . . 544 

244. Nor would the higher purposes of marriage be served by 

making infidelity penal, for they depend on disposition, 

not on outward acts or forbearances . , . .546 

245. All that the state can do, therefore, is to make divorce for 

adultery easy, and to make marriage as serious a matter 

as possible ......... 547 

246. (b) Should divorce be allowed except for adultery ? Some- 

times for lunacy or cruelty, but not for incompatibility, the 
object of the state being to make marriage a ' consortium 
omnisvitse' 547 



P. Rights and virtues. 

247. Outline of remaining lectures, on (1) rights connected with 

the functions of government, (2) social virtues. (The 
antithesis of ' social ' and ' self- regarding ' is false) . . 550 

248. Virtues, being dispositions to exercise rights, are best co- 

ordinated with rights. Thus to the right of life corre- 
spond those virtues which maintain life against nature, 
force, and animal passion . . . . . .550 

Similarly there are active virtues, corresponding to the 
negative obligations imposed by property and marriage . 551 

' Moral sentiments ' should be classified with the virtues, of 
which they are weaker forms . . ... . . 552 

251. Although for clearness obligations must be treated apart 
from moral duties, they are really the outer and inner side 
of one spiritual development, in the joint result of which 
the idea of perfection is fulfilled . . . . . 552 




P. 96, for ■ (Jr. A. Lewes ' read < G. H. Lewes.' 

P. 307, add to the Note of the Editor, ' See Prolegomena to Ethics, Book ii., 
eh. i., sec. 100, Editor's note.' 



Note of the Editor. 

The following lectures on Kant's Critique of Pure Benson were not all 
composed at the same time. A, B, F, G are extracts from a continuous 
course which was delivered more than once when Green was a tutor at 
Balliol College. They were written after the publication of the Introduc- 
tions to Hume (1874), but not long after, as they are referred to in lectures 
on logic delivered in 1874-1875. C, D, E, H, I seem to have been later 
additions or supplements to the previous course, to which they sometimes 
refer as 'old.' They were all apparently used, in whole or in part, for 
lectures delivered in Balliol College in 1875-1876. References show that 
at the time of their composition the articles on H. Spencer were already 
written, though the first of these was not published until December 1877. 

The lectures are here arranged in the order of the parts of the Critique 
to which they chiefly relate. The nature of the materials from which they 
are taken made some repetitions unavoidable ; and the abruptness of the 
transitions, and the occasional variation of view which they exhibit, are due 
to the same cause. 

The references are to the pages in Hartenstein's edition of Kant's works, 
vol. iv. The translations referred to are those of J. M. D. Meiklejohn and 
(in C) J. P. Mahaffy. 





1. Kant's speculation is governed by three leading 
notions : (a) that of an ' object affecting the senses ' and 
6 mind,' as independent existences, each contributing so much 
to knowledge; (how much, has to be settled); (6) that the 
c universality and necessity ' of a judgment are tests of a 
judgment being the work of mind, not the representation of 
an ' affection of the senses ' ; (c) that judgments may be 
divided, as into two mutually exclusive classes, into such as 
' merely analyse a subject- conception into its constituent 
conceptions, which were already thought in it, though in a 
confused manner,' and such as ' add to it some attribute that 
lies completely out of it,' of which classes the latter only 
represent an c augmentation ' of knowledge ; and that there- 
fore judgments which represent not only the work of 
c reason,' but its work as contributing to knowledge, must 
be ( synthetical,' not c analytical.' 

These notions of Kant come to him through Locke and 
Hume. His i objects affecting the senses ' correspond to 
Locke's ' substances that operate on us whether we will or 
no ' ; his ' empirical knowledge and judgments ' to Locke's 
' knowledge ' or ' propositions concerning substances or co- 
existence ' ; his distinction of ' synthetical ' and ' analytical ' 
to Locke's distinction of ' instructive ' and ' trifling ' proposi- 

2. The way in which Kant arrives at his conclusions 
about a priori knowledge is best understood by considering 


the difficulty into which Locke got with regard to mathe- 
matics and physical science by starting with the notion that 
all knowledge is the result of ' substances operating ' on the 
blank tablet of the mind. 'General certainty is never to 
be found but in our ideas.' The correlative of this is that 
the ' science of nature is impossible.' l Locke comes to this 
conclusion as follows : ' The operation of a substance upon 
us ' is an event in the way of feeling, a simple idea which 
can only be represented by a singular proposition : * This 
rose is red,' &c. We may retain these events in memory, 
and put together the simple ideas into complex ones, repre- 
sented by common nouns. These common nouns again may 
be made subjects of universal propositions, but such proposi- 
tions are either compendia of a number of past events (once 
and again and again I have found gold soluble in aqua 
regia), and are thus not properly universal, or, if properly 
universal, they relate merely to ' a nominal essence,' are 
analyses of the meaning of a name. 

Thus all ' propositions concerning substances,' it would 
seem, must be either e trifling ' or singular. In fact, how- 
ever, in speaking of a proposition as ' concerning a substance,' 
we imply that it is not ' trifling,' not analytical of the mean- 
ing of a name. Such a proposition as ' all gold is soluble 
in aqua regia ' implies (as Kant would say) the synthesis of 
solubility with a conception of gold which does not already 
include it. But thus understood — as ' instructive ' or syn- 
thetical — it has not the certainty which would belong to it 
if it were ' trifling ' or analytical, e since we can never, from 
the consideration of the ideas themselves, with certainty 
affirm ' their coexistence. 2 

Thus all propositions concerning substances (as = syn- 
thetical propositions derived from experience), if general, 
can only be problematical. 

What is the distinction between the relation of ideas 
expressed by mathematical propositions and the coexistence 
of properties (also a relation of ideas) expressed by ' all gold 
is soluble in aqua regia ' ? Locke's answer is that the ideas, 
of which the relation is expressed by the latter proposition, 
are ' ectypes ' of an e archetype.' There is a thing which 

1 Locke, Essay on Human Under- 2 Locke, loc. cit. Book iv. chap. vi. 

standing, Book iv. chap. vi. sec. 1 6, and sec. 9, and cf. General Introduction to 
Book iv. chap. xii. sec. 10. Hume, § 122, vol. i. p. 103. 

m 2 


produces certain ideas in us, and the proposition in question 
professes to state a relation between ideas as products of 
such a thing ; it expresses that there is uniform connection 
in a thing of that which the idea of solubility represents, 
and that which malleability &c. represent. And this is 
what we cannot know. We can only know that now or then 
the one idea has been produced along with the other. In 
mathematics, however, there is no ' archetype ' in question. 
Our mathematical ideas are themselves originals, and there- 
fore a relation between them that obtains once obtains neces- 
sarily always. 1 

First, then, we have the question how a mind, either 
merely passive or active only in the way of compounding 
and abstracting, can originate * instructive propositions.' 

Next, the question how these propositions can be really 
true, if reality consists in what the mind passively receives, 
in distinction from what it does for itself. 

Next, these propositions, which are really and universally 
true because they represent relations of ideas which have no 
archetypes other than themselves, are propositions about 
number and magnitude ; i.e. about ' primary qualities of 
body,' of which it is the differentia according to Locke, that, 
though our perceptions are copies of them, yet they are in 
things quite independently of our ideas of them ('whether 
there is a mind to perceive them or no ' 

Thus accepting the antithesis (which Kant retains from 
Locke) between what the object or thing gives and what the 
mind does for itself, Locke himself, in order to explain the 
nature of mathematical truths, has against himself to admit 
that they represent the original workmanship of mind. The 
antithesis itself, however, labours under great difficulties in 
Locke. All relations are the e workmanship of mind.' (This 
is a necessary admission because no relation is a simple idea, 
an event in the way of feeling.) But remove relations, and 
what is left of the object? ' Something whieh causes our 
feelings.' But its causation of our feelings is a relation. 
'Something merely.' But this ' something' represents the 
abstraction of difference, difference of which you cannot say 
in what it consists. 

3. Hume's problem is to render the doctrine of the 
mind's mere passivity consistent with itself by getting rid 

1 See General Introduction to Hume, §§ 116, 117, vol. i. pp. 95-97. 


of the relations which the mind constitutes, in particular 
mathematical relations, and those of cause and substance 
(the unity of successive appearances in an identical thing) . 
These are to be reduced either to the succession of one 
feeling on another, or to habitual propensities produced by 
the repetition of such succession. 

There is no idea not copied from an impression, i.e. from 
a feeling which carries with it no reference to anything other 
than itself. The distinction between mind and thing disap- 
pears in succession of feeling. Thus Locke's distinction of 
mathematical relations, as relations of ideas in the mind, 
from the coexistence of sensible qualities in a substance, 
cannot be maintained. The ' idea of space ' is a copy of the 
impression of space, and the impression of space is merely a 
compound of other simple feelings which come to us like 
any other feelings, we know not how or why. 1 I will not 
criticise this account at present. The point is to see how 
in Hume's judgment it affects the universality of mathe- 
matical truths. 2 According to him, when the mathematician 
talks of certain angles as always equal, of certain lines as 
never meeting, he is either making statements that are 
untrue or speaking of nonentities, of which he has a ten- 
dency to suppose the existence. But this * tendency to sup- 
pose ' is merely a way of saving appearances. If that which 
is ' supposed ' is neither impression nor idea, it is nothing. 

Thus with Hume there are no synthetical propositions in 
geometry that are universally true. He admits, however, 
that such propositions are possible with regard to number ; 3 
and this corresponds to his virtual admission that ideas of 
number are not copied from impressions. 4 

4. The notion that the object, given independently of 
thought, contributes so much to knowledge, viz. sensations, 
and that the subject on its side adds so much of its own, viz. 
relations or forms, is the basis of the doctrine that you can 
divide truths into such as are particular and contingent on 
the one side, and such as are universal and necessary on the 

Against this it is to be maintained that without thought 
there is no object; that feelings derive only from relation to 
thought that character in virtue of which we oppose them 

1 General Introduction to Hume, vol.i. * lh. p. 231, ff. 

pp. 202 and 203. 3 lb. p. 234. 4 77> p. 223. 


to thought ; that thus it is as impossible to divide knowledge 
into elements, one contributed by feeling, the other by 
thought, as to analyse the life of an animal into so much 
resulting from the action of the lungs, so much from the 
action of the heart ; and accordingly equally impossible to 
divide judgments into empirical and a priori. If ' experience ' 
means mere succession of feelings, then it yields no judg- 
ments at all, particular ones as little as universal. If it 
means ' experience of objects ' — such as can alone yield even 
singular propositions — then there is no ground for dividing 
truths into ' contingent,' derived from experience, and 
' necessary,' not so derived. 

There are two ways of understanding 'object.' 'Object' 
means either ' a permanent possibility of sensation,' and a 
particular object a permanent possibility of a certain set of 
sensations, or it means a * thing in itself.' ' A permanent 
possibility of sensation ' is not actually any sensation. No 
succession of feelings, apart from reference to a subject 
present to the succession but not in it, and determining the 
succession by distinction from itself, could be, or give the 
notion of, such possibility. Thus it is only through relation 
to a self-conscious subject that feelings are related to an 
' object ' in the above sense. Further, only as related to such 
an object, and thus becoming coexistent qualities instead of 
a mere succession of which one has ceased before the next 
begins, have feelings any character, anything that can be 
represented by a proposition. 

Thus the singular judgment, as the determination of an 
object by a certain relation, or the reference of feeling as an 
attribute to an object, is already the ' work of thought.' 

' Gold is soluble in aqua regia ' is supposed to be a con- 
tingent truth ; i.e. to represent certain events in the way of 
feeling, on the continued recurrence of which its truth is 
contingent. In fact, however, the scientific man does not 
treat it as contingent in this sense at all. He goes on the 
principle that whatever is really thus soluble in a single 
instance is so always, and accordingly, if he found a case 
where it seemed not soluble, he would decide, not that what 
had really happened in the previous case had ceased to 
happen, but that the conditions were different. 

In other words, he regards the singular proposition, 
' this gold is being solved,' as representing a universal law, 


and thus as a proposition which might be immediately con- 
verted into a universal proposition, if all the properties of 
gold and aqua regia were known and it were agreed to call 
nothing gold or aqua regia which had not th em- 
it is true, then, that experience constitutes universal 
judgments, but it does so just because the experience which 
constitutes singular judgments is already more than the 
occurrence of a feeling ; because it is an experience of an 
object regarded as always remaining the same under the 
same conditions. The modern logician is quite entitled to 
say (as against Kant) that experience can yield universal 
synthetical propositions, but not entitled at the same time 
to retain Locke's view of a singular proposition, viz. that it 
represents merely a sensible event or feeling. 



5. Two inconsistent notions struggle with each other in 

(a) The notion of ordinary logic, that ' things ' or ' ob- 
jects ' are given independently of thought, but having certain 
relations to our ' sensory ' which = sensible qualities ; that 
thought proceeds to detach these from each other and from 
s circumstances of time and place, 5 as abstract ideas or con- 
ceptions through which (represented by a common noun) 
henceforth it is 'mediately related to the object.' Mean- 
while the ' thing in itself (or, as Locke said, the ' thing in its 
real essence ' as distinguished from the ' nominal essence ') 
remains wholly unknown. It produces appearances from 
which we abstract a notion of it, but these, being all con- 
ditioned by the subject, tell us nothing of the real nature of 
the thing. When I experience a sensation, I can judge 
certainly " the thing is now affecting me,' but what is retained 
in the mind as the result of the affection and put together 
with other like results into a conception of the thing, is not a 
quality that belongs to the thing ; it does not enable me to 
assert anything really true of the thing. 

From this notion arises (1) his antithesis of analytical 
and synthetical propositions. A proposition, to be instructive 
or convey information about matter of fact, must relate to a 
real object, as opposed to a conception. (2) In like manner 
(unless its universality can be explained on some other 
ground) it can only be particular and contingent, because 
the ' thing in itself,' some effect of which it represents, may 
be pleased to produce another effect at another time. (In 
truth, the antithesis will not hold, for if we remove from the 
instructive proposition that content of its subject or predi- 
cate which is merely conceived, no meaning is left.) 

(b) On the other hand, there is in Kant the notion that 


the object first becomes an object through a certain action 
of the mind upon a ' matter ' given in sensation, and a real 
object through imposition on this matter of ' forms of intui- 
tion.' Form of intuition being the condition of reality, 
truths relating to it will be ' instructive.' They will have 
that privilege, supposed in the other theory to belong to truth 
concerning real objects, without the drawback of being merely 

Admitting this notion, the question naturally arises, (1) 
what meaning there is in talking any longer of a ' thing in 
itself at all, when the ' object,' which according to (a) was a 
' thing in itself,' has turned out not to be independent of 
mind; (2) whether there is any such distinction between 
* intuition ' and 4 conception ' as can take the place (which 
Kant seems to give it) of the distinction between what is real 
and what is of the mind, implied in notion (a). 

6. It is quite true that space and time are not ' relations 
of objects' as opposed to ideas, of which afterwards by 
abstraction there come to be ideas (conceptions). They are 
themselves ' ideas,' which are the condition of there being 
any phenomenal object whose qualities may be abstracted, 
the condition of mere feeling becoming a felt thing. But 
just for this reason, though ideas, they are relations of 
objects ; and we are confusedly conscious of the object as 
conditioned by these relations before we think separately of 
these conditions themselves. 

It is commonly thought (a) that to admit space and time 
to be relations of objects (real things) is incompatible with 
their being in any sense ideas (though there may be ideas 
o/them), and (b) that to admit that we, as learning indivi- 
duals, have ideas of successive objects, and objects outside 
each other, before we have ideas of space and time as such, is 
incompatible with the a priori character of the latter ideas. 

But why should 'relations of objects' not be ideas? 
We fancy that they cannot because our notion of an idea is 
just that it is not a reality : it is determined simply as the 
negation of a thing. This opposition arises from the fact 
that our ideas change through the operation of an experience 
which we do not make. But this fact need not imply that 
there is any reality other than ideas, but only that ideas are 
communicated to us gradually : and when we reflect that 
our existing ideas at any time qualify all new experience 


(which only derives meaning from relation to them), it will 
appear that the progressive character of our knowledge is 
better explained as a revelation of the actually existing 
ideas through which possibilities of them in us are gradually 
actualised, than as the result of an operation of things, 
which are not ideas, on us. Our experience then does not 
require the supposition of such 'things,' and after all, what 
are they ? They are either subjects of the qualities which 
make up our experience, or 'things in themselves.' Now 
the * qualities which make up our experience ' mean relations 
between feelings constituted by the presence to the feelings of 
a self-conscious subject. 'Things,' then, the supposed oppo- 
site of a thinking consciousness, if they are the ' subjects of 
the qualities which make up our experience,' are determined 
by, are what they are in virtue of, a thinking consciousness. 
Nor less is the ' thing in itself ' determined by our thinking 
consciousness, though determined as its negation. If, then, 
the opposition between idea and reality will not hold good, 
the meaning which we seem to derive for ' idea ' from this 
opposition turns out to be none at all. What then is an 
' idea ' ? It is a community between objects, which is at 
the same time their difference, a community constituted by 
the presence of a single self-conscious subject to the mani- 
fold of feeling. In this sense, space and time are at once 
' ideas ' and ' relations of real things.' Their characteristics 
as relations are (1) primariness, from which it arises that 
whatever is true about them is so unconditionally ; (2) 
simplicity, from which follows the ease of ascertaining pre- 
cisely what is true of figures. If we could ascertain any 
truth about (say) the relations of chemical substances in the 
same exact and unconditional way, it would be equally 

7. This primariness or a prior i character of the ideas 
which constitute space and time is not to be understood as 
priority in time, as if we had the idea of mere space before 
any other ideas. There is no experience of space apart from 
colour and tangibility, nor do we present mere space to our- 
selves before such experience. The primariness of the idea 
means that it is the condition, without which no feelings 
would become outward things, so that all other conditions 
of ' phsenomena ' may be supposed absent, but not that. 
Hence it is that we can present to ourselves things as having 


no other properties but what arise from this relation ; i.e. as 
spaces and nothing else, as mere spaces. In this lies the 
explanation of Kant's distinction between the idea of space 
as an intuition and other ideas as conceptions. The objects 
of which space is predicable are parts of space, and there is 
no more or less in one space than in another, or in any part 
of space than in space in general. (This, however, is only true 
of pure space.) l In other words, space is space just the same, 
irrespectively of what it contains, and space is a quantum, 
of which, as of every other quantum, it is true that each of 
its parts is also a quantum, i.e. that it is infinitely divisible. 
This is because from the primariness of the relation of out- 
sideness we can present to ourselves objects as determined 
by it, and by no other (i.e. as mere spaces), and can put these 
together as a quantum of which all the parts are homogeneous 
with each other and the whole. 

8. We have here undoubtedly a peculiarity in the ' idea ' 
in which consists the relation of space. The objects between 
which subsist the relations expressed by chemical affinity 
cannot be regarded as qualified no otherwise than by those 
relations ; for the relations, so soon as you think of them, 
branch out into others, complex as the universe. For that 
reason ' humanity ' cannot be presented as a quantum, of 
which the individuals related in the way of humanity are 
parts. But is this difference any reason for questioning 
that space and chemical affinity are alike conceived re- 
lations, not relations that first exist and then are conceived, 
but relations constituted by the presence of a single self-con- 
scious subject to the manifold of feeling ? 

6 Intuition ' with Kant is the presentation of a real indi- 
vidual object ('real' as phenomenon) . 'Conception ' is the 
thought of an attribute or attributes possessed by such an 
object. To all knowledge (to all judgment representing 
knowledge) ' intuition ' is necessary, though not to a mere 
logical or analytical judgment, which unfolds the content of 
a conception. 

1 ' Is there,' it may be asked, ' really does not depend on any other relations, 

such a thing as pure space?' Not if It may therefore be considered sepa- 

really means ' fur sich bestehend ' ; there rately and presented as an object, about 

is nothing ' f iir sich bestehend' but which judgments may be formed which 

thought if self. There actually is such a will at once be true and will relate to 

relation as that of the limit or exter- matters of fact, 
nality, and the nature of this relation 


The idea of space is the presentation of a real individual 
object: the idea of humanity is not. The question then is, 
what beyond relations is expressed by ' object/ * individual,' 
and ' real ' ? 1 Object ' expresses a relation to consciousness, 
a relation in which each of the constituents of the relation 
is determined by opposition to the other. ' Individuality ' 
represents distinctness of an object from all others, a distinct- 
ness constituted by the complex of its relations, or, if it 
merely be separateness in space and time, still by relations. 
' Eeal ' represents the identification of the object, here and 
now given, with previous presentations. (I seem to see a 
horse : is it a real horse or a delusion ? i.e. are the relations 
of the object now before me the same as those in virtue ot 
which I have denominated objects previously presented as 
' horse'?) If we choose to mean by conception the fixing 
under a name of some particular relation or relations apart 
from others, then doubtless the distinction is valid between 
it and real individual things ; not, however, as a distinction 
between relations constituted by thought and anything other 
than they, but as one between an isolated set of relations 
which we first learn to know and those with which the pro- 
gressive communication of thought to us is gradually making 
us acquainted. 



[Krit. d. r. V. pp. 565-585. 1 ] 

9. The c transcendental deduction of conceptions ' = the 
' explanation of the manner in which conceptions can relate 
a priori to objects.' 

But why assume that there are such conceptions ? why 
not suppose that all conceptions are derived from objects 
a posteriori by experience, through abstraction and general- 
isation 9 

Kant's answer would be that there are certain conceptions 
which are necessary in order to render objects of experience 
(objects as connected in a world of consciousness) possible, 
which therefore cannot be derived from them. His views on 
this point are the first thing we have to consider. We shall 
then come to his ' deduction ' in the sense explained above. 
Characteristic of this is his view that conceptions could not 
' relate a priori to objects,' if the objects were ' things in 
themselves.' It may be asked, indeed, how, if pure concep- 
tions are necessary to render objects of experience possible, 
any question can be raised as to the possibility of their 
relating to objects. Kant, however, speaks of 6 objects ' in two 
different ways. Objects which pure conceptions render pos- 
sible are objects as connected in the ' cosmos of experience.' 
Objects, as to which he asks how pure conceptions can relate 
to them a priori, are objects as not yet connected in such a 
cosmos. In his language they are objects of intuition. If 
these were ' things in themselves,' i.e. other than our repre- 
sentations, pure conceptions (being of subjective origin) 
could not relate to them. In fact, they are affections of our 
sensibility, produced indeed (such is Kant's view) by 'things 

1 [The translation referred to in this section is that in J. P. Mahaffy's Kent's 
Critical Philosophy for English Readers, vol. iii.] 


in themselves ' of which, we know nothing, but determined by 
a priori forms of our sensibility, space and time. It is 
because objects of intuition are of such a sort— determina- 
tions of the affections of our sensibility by the subjective forms 
of space and time — that pure conceptions can relate to them 
a priori, and out of them construct the connected whole of 
our experience. 

10. Let us now return to the first question, What ground 
is there for holding that there are i pure conceptions which 
relate a priori to objects ' ? Kant's doctrine generally 
seems a laborious effort to meet a difficulty which does not 
exist, because it is understood as dealing merely with our 
conceptions of certain relations, not with those relations 
themselves. Now the essence of Kant's doctrine is that it 
deals with the relations themselves. Our experience consists 
of related phenomena, i.e. related feelings. His question 
is, How comes it that feelings thus form an inter-related 
whole ? and it is not met by a doctrine which, taking the 
relations for granted, traces the process by which we become 
certain that the propositions which represent them are 
universally true. Kant's point is that only the act or pro- 
cess of conception constitutes these relations, and that it is 
preposterous to derive such conception from the experience 
which, by thus constituting uniform relations, it renders 
possible. A Kantian may fairly be called on to explain how it 
is that the conceptions, which have been necessary in order to 
constitute the relations of which we have experience, only 
come into distinct consciousness after a long course of experi- 
ence ; but a theory as to the origin of certain relations is 
not answered by one which takes them for granted. 

To take a particular instance. All * psychological ' 
accounts of the origin of the conception of cause presuppose 
relations of identity, change, and succession. The simplest 
form of experience which is supposed to suggest the concep- 
tion is somewhat as follows : — an object, supposed to retain 
its identity, to be the same as when we had previous expe- 
rience of it, is yet found to have undergone a change, to 
appear different in some respect from what it did before. 
This excites surprise, and suggests inquiry how the change 
comes about. It is observed to occur uniformly in succession 
upon some other appearance, and thus we associate the two 
appearances in our mind as cause and effect. 


What is implied in there being, how comes it about that 
there are, these relations of succession, change, and identity 
as relations between phenomena, or for consciousness ? that 
the experience of mankind forms a connected whole in which 
variations have to be explained as consistent with the uni- 
formity of the whole ? 

Events, unrelated, could not be a succession. The possi- 
bility of a succession implies something other than the 
things which succeed. In order to their being related even 
in the way of sequence, there must be some unit, other than 
the events, and not passing with them, through relation to 
which they are related to each other : a, b, c are points in 
succession ; a is over when b begins, otherwise they are not 
successive. There must be something else, then, for which a 
is not over when b begins — for which it is still present — in 
order that the two may be related to each other as present 
to past. 

11. Need this unit, through relation to which events are 
related to each other, be conscious ? Granted that any succes- 
sion implies a relatively permanent something through succes- 
sive relation to which events are successive upon each other, 
need this be other than a prolonged event, e.g. an organic 
body, which is the subject of perpetual changes, continuing 
throughout them, and yet itself passes away ; which is thus 
permanent relatively to the events which take place in it 
without disintegration of the organism, but not absolutely so ? 
The answer is, that just so far as it is not absolutely so, 
something else is implied ; that is, some unit through common 
relation to which the organism and other events before and 
after it are related in the way of succession. Still, must 
the ultimate unit be conscious ? may it not be an unchange- 
able matter, a sum of atoms which remains the same through 
all the changes of their distribution which constitute the 
history of the universe ? 

The answer is that, if you suppose an ultimate uncon- 
scious unit, you still require a further conscious unit to 
correlate the unconscious unit with the manifold events 
which, through relation to it, are related to each other. 

We speak, it is true, of unconscious agents, forces, com- 
bining manifold materials. Such combination, however, 
has nothing in common with the constitution of a relation 
between events. The agent or force, to which combining 


power is ascribed, is really a name for the relation between 
certain events, or for the conditions nnder which one follows 
another, e.g. one combination of chemical elements takes 
the place of another. An agent or force, thus reducible to 
some mode of relation of events to events, is quite different 
from that which is required to render such relation possible. 
All nature may be said to consist in the action of the uncon- 
scious npon the unconscious, but the unconsciousness of the 
factors to a relation must not be confused with unconscious- 
ness on the part of the correlating unit. In calling the 
relation a ' phenomenon ' we have said that it is for conscious- 
ness that it exists, and — not to argue from a term which 
people use without much meaning — what is that which 
retains a plurality in its plurality, and yet unifies it through 
relation, but consciousness? We know consciousness as 
that to which the past is yet present, and present as past ; 
in which a manifold is united in one experience without 
ceasing to be manifold. Why seek another source of rela- 
tion for the cosmos when we have that which suffices for the 
work? Why seek it in 'unconscious matter,' which after 
all means nothing but the bare negation of that which alone 
we know of as serving the pxirpose for which this ' uncon- 
scious matter ' is assumed? 

12. So much by way of preliminary. Now let us consider 
in detail Kant's account of that unity of consciousness which 
is also the unity of the world. 

In the simplest knowledge of an object there is involved 
a threefold synthesis, that of apprehension, that of imagina- 
tion, that of recognition. 

' Every intuition contains in itself a multiplicity, which 
nevertheless would not be represented as snch, if the mind 
did not distinguish time in the sequence of impressions one 
upon another ; for, so far as it is contained in a single instant, 
no representation could ever be anything but an absolute 
unity.' ] 

The intuition, this or that individual object as presented 
here and now, 'contains multiplicity,' i.e. is made np of 
parts. The intuited object really does so, but intuition is a 
single act, not multiplex. How then is the intuited object, 
as the representation of the object of consciousness, mani- 
fold ? ■ Because we distinguish time in the sequence of 

1 P. 567, Tr. p. 194. 


impressions.' I.e. (if I understand) because the object is 
given to us through successive acts of attention. The data 
of these several acts we then hold together, without fusion 
with each other, as parts of one whole. 

Feelings may follow upon each other, but only in relation 
to a subject equally present to each — for which each is not 
over when the other begins — do they form a succession or is 
there a distinction of time between them. It is the synthetic 
act of this subject which renders certain sequent impressions 
a manifold in one, an individual or ' intuited ' object having 
variety of parts or qualities. 1 

' Intuition ' = perceived object. This has become a mani- 
fold for consciousness or been represented as manifold because 
of the multiplicity of times in which it is attended to. 
This multiplicity is at once ' run through and grasped 
together,' and there thus results a many-in-one, a complex 
object. A complex object having been thus constituted for 
consciousness, a 'representation contained in a single in- 
stant ' suffices to recall it. 

In the last clause of the paragraph ' intuition ' seems to 
be used as equivalent to attention, or rather to the several data 
of successive acts of attention upon which the ' synthesis of 
apprehension is directed.' These are many, but only form a 
manifold in one representation in virtue of the operation of a 
synthetic principle in and with successive acts of attention. 

13. Hereupon arise certain questions. (1) What is the 
'intuition' ( = perceived object) in itself, which is said to 
contain a multiplicity, as distinct from the representation of 
that multiplicity ? Like every object it is the possibility of 
certain perceptions occurring in a certain connection. It 
contains multiplicity in itself, only as the possibility of cer- 
tain experiences on our part ; the reality is those experiences, 
as connected. An ' intuition ' or individual thing is really 
nothing apart from these, i.e. it is really nothing save as 
constituted by the synthetic act described. (2) Why speak as 
if the multiplicity of the intuition as represented was merely 
a multiplicity of times (arising from ' the distinction of time 
in the sequence of impressions ') ? As we have seen, there 
would be no multiplicity in representation but for the data 

1 This will become clearer if we con- p. 414. Mr. Spencer on the Independ- 
sider the matter without the use of tnce of Matter, sec. 37 . 
Kant's technical language. See vol. i. 

VOL. II. n 


of successive acts of attention and the synthesis of these, bnt 
these are not merely times, though given in successive times ; 
if they were, the intuited or perceived object would be time. 
Kant, however, wants to show that there is a synthesis of 
apprehension which is quite 'pure.' Apparently it is not 
enough for him that it should be ' pure ' in the sense that an 
agent other than feeling is necessary to constitute it accord- 
ing to the process which we have described ; necessary, (a) as 
distinguishing feeling from itself in the act of attending to 
it, (b) as holding the data of successive acts of attention 
together in virtue of equal presence to each. He requires a 
synthesis which is 'pure' in the sense of constituting an 
object out of the pure forms of sensibility, a synthesis 
exercised by understanding upon material consisting of mere 
distinctions of time and space. These distinctions = < the 
manifold, which sensibility offers in its original receptivity.' l 

Peelings, through the action of the unity of the under- 
standing upon them, are related in the way of succession ; 
they occur in manifold times which (in virtue of the same 
unity) are held together as one object, a time made up of 
homogeneous parts or times. Such a representation of time, 
according to Kant, though constituted by the unity of under- 
standing, is not a conception but an intuition. He confines 
the term conception to the thought of attributes or relations 
common to several individual objects, as opposed to the 
representation of individual objects, of which we say ' this ' 
or ' that.' But of time we can say ' this ' or ' that,' and 
whereas of ' horse in general ' or ' horse as conceived ' we 
cannot say all that we can say of this or that horse, of ' time 
in general' we can say just the same as of this or that 
time. Time, then, is an intuition, not a conception. That 
unity of a manifold which constitutes time is not a unity 
abstracted from many intuited objects, but is a unity neces- 
sary to constitute this or that time, and thus belongs to 
time as an object, not to our abstract conception of it. 

According to Kant, the action of the unity of the under- 

1 Sensibility, according to Kant's relations' (Trans. Msth. sec. 1); that 

usage, does not equal sensation, but which makes the difference between 

'perception. Original receptivity seems sensation and perception. For his 

here to be used for what he elsewhere view of the way in which ' formal in- 

calls the form, as opposed to the tuitions ' are constructed out of pure 

matter, of sensation, viz. 'that which 'forms of intuition' see Krit. d. r. V., 

effects that the content of the pheno- p. 132, note; p. 98, Tr. 
menon can be arranged under certain 


standing upon the manifold given in the outer sense yields 
the pure intuition of space in the same way as its action 
upon the manifold given in the inner sense yields the intui- 
tion of time. The representation of space is not a conception, 
because of space in general everything can be said which can 
be said of this or that space. 

The question of the distinction between outer and inner 
is a great difficulty, which must be postponed. 1 The great 
point to bear in mind for the present is that the unity of 
understanding is just as necessary in order to yield the 
representation of a manifold as to unify that manifold, because 
in order to the representation of a manifold not only must 
successive feelings occur, but some synthetic subject must 
carry the fact that the first feeling has occurred on with it 
to the second and third, and so on. Synthesis of apprehen- 
sion, then, is that which is necessary to constitute any per- 
ception of an object, 2 of this or that, since the object of 
perception always is individual. 

Such synthesis, even in the case of ' empirical intuition/ 
must contain a e pure element,' the correlating principle, but 
Kant does not call it ' pure ' except when exercised upon 
pure forms of sensibility (i.e. the conditions of distinctness 
in space under which the data of outer sense, of distinctness 
in time under which the data of inner sense, are presented 
to us), and so constituting a c formal intuition ' of this or 
that space, this or that time, or of pure space and pure time. 

14. The ' synthesis of apprehension ' involves that of ' re- 
production.' The qualities which we combine in one object 
of perception are really relations to past sensations (relations 
not being themselves past), and the synthesis of apprehension 
would be impossible unless we could recall experiences in 
which these relations were given. There must, moreover, 
have been some * rule ' according to which these experiences 
were connected together. Otherwise, though we might have 

1 [See below, § 54, ff.] objects of perception. This ' thing in 

2 It is a matter of indifference itself Kant regarded as 'rendering 
whether you say perception of object or nature possible on the material side,' 
perceived object, for it is the synthesis just as ' unity of apperception ' ren- 
implied in successive perceptions and ders it possible 'on the formal side.' 
the synthesis of these syntheses that {Prolegomena zu ciner jedcn kunftigen 
constitutes the object. The object in Metaphysik, sec. 36.) The question 
itself must mean either the abstraction is whether it is really more tnan the 
of the possibility of these, or the • thing abstraction just mentioned of the pos- 
in itself as the unknown cause of these sibility of experience, as distinct from 
sensations, relations between which are its reality. 

c 2 


the power of recalling them, there would be nothing to 
make us recall any particular series of them, on occasion of 
a present sensation, as representing qualities of the same 
sensible thing which the sensation represents. This ' rule * 
we might suppose to lie in some 'thing in itself,' if that 
which it connects were anything else than phenomena. As 
it is, it must belong to the same consciousness to which 
phenomena belong. Thus Kant calls it the ' transcendental 
unity of imagination.' 

It is not enough to say that I recall a and b on the re- 
currence of c because I have constantly had experience of a 
and b as immediately preceding c. The question is, how 
upon the first sequence of a, b, and c I was able to carry on 
the experience of a and b into the experience of c so as to 
connect them in one experience. It must have been in virtue 
of * an a priori ground of necessary synthetic unity,' a e syn- 
thesis of representation ' not e empirical,' i.e. not gradually 
resulting from experience, but ' transcendental,' i.e. which 
renders experience possible. And this synthetic principle, 
which originally determined the connection of certain suc- 
cessive experiences with each other, is really the ( rule ' 
which on the recurrence of one of these experiences deter- 
mines the recurrence, the representation, of the rest in a 
necessary order (so that if the order of original experiences 
has been a b, c, d, on the recurrence of d, a could only be 
recalled through c and b). If we ask what this 'a priori 
ground ' is, we shall find (though he does not say so here) 
that according to Kant it is just the one subject present to 
all experiences. 

At first it strikes us as a superfluous question to ask how 
it is that, on the occasion of a certain experience in the way 
of feeling, other experiences are recalled in a determinate 
order. It seems enough to say that it is because they have 
constantly occurred in this order. ' But how has this order 
been possible ? ' ' This is a still more superfluous question. 
It is an order determined by nature.' ' Just so, but what 
is nature P ' If nature = a ' thing in itself,' the unknown 
opposite of our representations or consciousness, it does not 
explain what has to be explained, which is just a law of 
consciousness, viz. how our representations are woven to- 
gether into one order ; how it is that not only feelings a, b, 
c d come one after the other, but that the experience of a 


(the consciousness that a has occurred) is carried on as a 
determinant of b, that of a and b having occurred as a 
determinant of c, and so on, so that there results a definite 
series in consciousness which can only be recalled in that 
precise order. What we really mean by nature, however, 
is not a • thing in itself,' but just this determinate order of 
phenomena or consciousness. To say, then, that such order 
is possible because it is the order of nature, is to say nothing. 
At any rate it is more to the purpose to point out, as Kant 
does, the condition of its possibility in the presence to all 
feelings of an identical subject, to which that which is no 
longer present as a feeling remains as the known fact that 
a feeling has occurred. 

Thus Kant's doctrine does not at all conflict with the 
psychology of association on the proper ground of that 
psychology. It is dealing with a previous question which 
that psychology either does not answer at all or answers by 
a tautology, the question how we are to explain that original 
connected consciousness, that first orderly experience of phe- 
nomena, which must have come into existence before it can 
be recalled in modes which the theory of association investi- 
gates ; the question, what are the conditions of its possibility. 

15. In order to the experience of an object — to any per- 
ception, in fact, as we commonly understand it — more than 
synthesis of ' intuition,' as including that of 6 reproduction,' 
is required, viz. synthesis of ' recognition.' 

Is synthesis of * intuition ' any more possible according to 
Kant's view without that of ' recognition ' than it is without 
that of ' reproduction ' ? 

On the one hand he constantly opposes 'intuition,' as 
that in which objects are given, to ' conception,' which takes 
them as given, and e synthesis of recognition ' he identifies 
with the act of conception. 

On the other hand it appears from the account of 
' synthesis of recognition ' that without it ' intuition ' would 
not be intuition of an object at all. The conclusion from 
this account would be that though i synthesis of intuition,' 
as the unification of a manifold resulting from successive 
acts of attention, might yield a mental image containing a 
multiplicity of parts, this image would not, without synthesis 
of recognition (act of conception), become an object which, on 
presentation of a like image, might be recognised as the same. 


But if so, how can ' intuition ' be said to give an object 
at all? We may try to make out that Kant speaks of 
'object' in two senses : (1) as an object presented here and 
now, which is said to be ' intuited ' ; (2) as an object connected, 
through the act of conception, with other objects in one 
system of experience. These two senses, however, are not 
really distinguishable, because the ' here and now ' imply 
' connection with other objects in one system of experience/ 
As individual, Kant reckons an object one of intuition, as 
related, one of conception ; but relations constitute individu- 
ality. The truth is that Kant's treatise is throughout per- 
plexed by his habit of accepting provisionally distinctions 
which it is the result of his work to invalidate. Thus he 
accepts that between intuition and conception. In the 
' ^Esthetic ' he speaks as if sensibility, in virtue of its pure 
forms, sufficed to yield intuitions of space and time. In the 
'Analytic' 1 he expressly points out that 'unity of under- 
standing ' is needed to constitute such intuitions, though he 
still seems to insist that they are intuitions, not conceptions. 
At the beginning of the sections on the ' synthesis of recog- 
nition ' he distinctly calls this at once a process of conception 
and a necessary condition of the knowledge of objects ; so 
that we must either say that the ' formal intuitions of space 
and time ' are not, as such, known objects, or that he con- 
tradicts himself in opposing them to conceptions. 

16. The ordinary account of perception is that a present 
sensation recalls sundry possibilities of sensation, which are 
referred to the same object as that to which the present 
sensation is referred. This implies that the present sensation 
is identified with one of which we have previously had 
experience in relation of antecedence, sequence, or simul- 
taneity to these sensations which we recall as possible ; and 
again that the latter, on their occurrence, have been identified 
in a similar way. The question, ' How is such identification 
possible ? ' is that which Kant deals with in his account of 
the synthesis of recognition. 2 

1 Pp. 131-132 ; p. 98, TV. with the facts of the previous occur- 

2 Such identification of a feeling is rence of such feeling. The degree to 
quite different from the mere reproduc- which recollection of these facts is ac- 
tion of it in a fainter form. It is not, companied by anything like reproduc- 
indeed, strictly speaking, the feeling tion of the past feelings, by the nervous 
which is identified with past feelings at modification implied in such reproduc- 
all, but the observed fact that a certain tion, varies indefinitely without affect- 
feeling is being felt which is identified ing the recollection. Probably there 


■ Without the consciousness that what we now think is 
identical with what we thought a moment ago, all repro- 
duction in the series of representations would be useless ' ; l 
i.e. unless the present experience were connected with that of 
a moment ago as related to the same object or as a change 
of the same subject; just as in counting, at each step units 
already traversed are carried on with him by the person 
counting to the next. Such connection is the identification 
of the fact of present sensation with the facts of past sensa- 
tion spoken of above. It is the condition alike of the 
simplest judgment e this is what that wa,s,' 'this is what I 
felt just now,' and of the scientific conception of nature as 
a system of which every part or process is determined by 
relation to all the rest. 

1 7. It is the condition, again, of there being for us such 
a thing as reality. ' I seemed to hear a voice, but I cannot 
really have done so.' What does this mean ? Ultimately it 
means that the reference of a certain feeling or impression, 
which I undoubtedly had, to such a vibration as in relation 
to the nerves of hearing constitutes the sound of a voice, is 
inconsistent with the necessity I am under of regarding all 
experience as a connected whole. 

We express the same thing, however, by saying that 
what I seemed to hear did not represent any real object. 
What do we mean by ' real object ' when we say so ? At 
first we are apt to suppose that there are a lot of separate 
things outside us. Then these things resolve themselves 
into certain possibilities of sensation, determined by an 
order of nature. This order, then, becomes the one object, 
apparently other than our consciousness, by reference to 
which we decide whether any interpretation of consciousness 
represents reality or no. 

It is in this sense that we speak of the ' object ' when we 
say, according to the definition which Kant probably had in 
his mind vvhen he wrote this section, that true knowledge is 
' agreement of thought with its object.' Now what can such 
an object be, of which, just because it is other than any 
determination of consciousness, nothing in particular can be 

need be no such reproduction. We are than actual feelings in perception, yet 

apt to confuse feeling as felt with the feelings are constituents of perception, 

observed fact of feeling, and thus to (See vol. i. p. 411, sec. 31.) 
supposo that though there may bo more ' P. 669 ; p. 197, Tr. 


said? It seems to be merely something in general, an un- 
known quantity, x. In truth it is the 'transcendental 
ground of the unity of consciousness in the synthesis of the 
manifold in all objects of experience ' * (' transcendental ' as 
that which conditions experience, instead of being con- 
ditioned by it). It renders possible all particular judgments 
about matters of fact, and just for that reason none of these 
judgments are applicable to it. 

It is other than consciousness, in the sense that it is not 
any or all of the particular modifications of consciousness. 
But it is that unity which binds all these into one system ; 
it is at once their real connection and the source in us of 
the progressive "knowledge of their connection. This ' pure, 
original, unchangeable consciousness 5 is what Kant calls 
4 transcendental apperception, 3 2 so called to distinguish it 
from 'empirical apperception' or 'internal sense,' which 
means consciousness of our successive states. No data of 
? empirical apperception ' are predicable of that ' numerically 
identical ' consciousness which renders ' empirical appercep- 
tion ' possible. Nothing is predicable of it but its function 
in constituting intelligent experience, or ' synthesis of recog- 
nition ' ; or, as Kant more particularly puts it, in compelling 
(a) reference of each 'representation,' as it occurs, to an 
object, so that it becomes a phenomenon or represented 
object, to which in turn other representations are referred; 
and (b) reference of all phenomena, thus determined and 
inter- related, to one ' object,' as representing which they are 
' real,' this ' object ' being in truth only another name for 
the unity constituted by the ' transcendental apperception,' 
the unchangeable thinking subject itself. 

18. Kant asks, ' What do we mean when we speak of the 
object corresponding to cognition ? ' The answer is : ' Some- 
thing in general, =x, because outside our cognition we surely 
possess nothing which we could place over against it, as 
corresponding to it.' 3 Does Kant, then, believe that there 
really is such an object ? Not exactly ; here, as elsewhere, 
he takes from the current theory of his time a doctrine which 
disappears under his hands. He inquires first what is its 
function in regard to knowledge. It is ' that which prevents 
our cognitions from being determined at random, or as we 
choose, but a priori in some certain way'; ('a priori 9 in 
1 P. 57! ; p. 200, Tr. 2 P. 572; p. 201, Tr. 8 P. 570; p. 198, Tr. 


the sense that the determination is not merely a habit result- 
ing from experience, but one that determines experience 
itself). The object, then, qua something other than all our 
representations, is for us nothing; while, in respect of what 
it does, it is the ground of synthesis in consciousness, 
' formal unity of consciousness, in the synthesis of the 
multiplicity of representations' (ib. p. 199). (Formal is 
opposed to material. Material unity of consciousness would 
be the unbroken continuance of the same feeling. Its formal 
unity is that which necessarily connects the most different 
and discontinuous feelings as one system.) This is the true 
meaning of ' object,' not only when we talk of knowledge in 
general as the agreement of thought with its object, but 
when we talk of knowing or perceiving (erhennen) this or 
that object. Such ' erkennen ' implies that a ' function of 
synthesis according to a rule ' has (a) formed successive 
feelings into a series, which in time determines the order of 
reproduction in memory or imagination (' makes the repro- 
duction of a manifold a priori necessary '), and (b) rendered 
possible the conception, * in which the reproduced manifold 
is united,' i.e. the conception of a relation between all the 
recalled facts of feeling (ib.) I see this or that object. 
This is an intuition according to Kant. We commonly say 
that it means that on occasion of a certain sensation I am 
aware of certain jjossible sensations, which would become 
actual if I did certain things, and which I regard as related 
to the actual sensation. This implies the reproduction of 
certain experiences and the conception of them as related. 
6 Object ' is merely the name for that which renders such 
reproduction and conception necessary, and that, says Kant, 
is the ' formal unity of consciousness,' in virtue of which 
feelings are (1) so related to each other as to form a series, 
reproducible in memory in one certain way only, and (2) 
conceived as related. 

The only fault to be found with Kant's account is that 
he speaks as if the constitution of a relation between feel- 
ings (which is what ' makes the reproduction of them a priori 
necessary'), and the conception of a relation between them, 
were two different synthetic functions of the ' formal unity 
of consciousness.' 

19. Kant illustrates by asking what we mpan by a triangle 
as an object, in distinction from the act of presenting a 


triangle to the mind's eye which we at any time perfor 
It means a relation, or (as he puts it) a certain mode of com- 
bining three straight lines according to a rule. (So any 
6 material ' object would be a mode in which certain sensa- 
tions follow or accompany each other according to a rule 
or definite relation between them.) The unity of the rule 
determines what elements of the c manifold,' i.e. what pre- 
sentations to sense, can be regarded as parts of a triangle 
combined in one ' total impression ' or intuition of a triangle. 
This conceived unity is really the object which we have 
before the mind when we make propositions about ' the 

Another illustration. The objects to which we refer the 
experience of outer sense we call ( bodies.' What does 
■ body ' mean ? Merely a rule according to which in intui- 
tion (i.e. perception), given certain phenomena, certain others 
are necessarily recalled and combined with them under cer- 
tain relations, such as extension, solidity, &c. 

The conception of such a rule cannot be of empirical 
origin. In however ' incomplete and obscure ' a form it 
must regulate experience, so as to make it experience of 
objects from the beginning. (This is so far quite true that 
all attempts to derive such formal conceptions as that of 
6 body ' from experience are found to treat the experience 
from which they derive it as already in some way an experi- 
ence of objects.) 

Kant's point becomes clearer if for ' conception of a rule ' 
we say conception of relation. No repetition of feelings can 
generate that conception of them as related, which makes 
us interpret one of them, on its occurrence, as one factor of 
relation of which the other must be in existence too, though 
no actual feeling represents it. 

Kant says ' there is always a transcendental condition at 
the foundation of any necessity.' This seems a needless 
assumption to those who take ' necessity ' to mean merely 
the strongest possible conviction of certainty. With Kant 
it means that which on ultimate analysis is found to be the 
condition of there being an object for us at all. In this case 
the necessity is that of referring intuitions to an object (or, 
as I think it would be more correct to say, of referring 
feelings to an object, so that there comes to be an intuition 
in the sense of perception). Without such reference there 


would be no experience. Its ground, then, must be * tran- 

20. It may be said, ' When Kant asks, what is the object 
corresponding to the intuition of a triangle ? what are the 
bodies to which we refer all experience of outer sense ? is 
not the true answer that each is a possibility? one the possi- 
bility of combining lines which we abstract from what we 
see or touch, the other the possibility of resistance?' Well, 
but does not such a possibility mean a conceived relation ? 
The possibility is a determinate possibility, not the possi- 
bility of chance. Body = possibility of resistance, i.e. the 
reference of experiences to bodies means the reference to 
something that would resist, i.e. require some expenditure 
of muscular effort before it would change place. Admitting 
this as a sufficient analysis of what is meant by body (which 
it is not), still this implies a necessary or objective relation 
between feelings (more than the fact that this has been felt 
and then that) ; it implies a law or relation which exists 
when the feelings are not being felt as much as when they 
are, according to which one can occur only in a definite 
relation to others. In order to the existence of such law or 
relation, there must be an eternal unit; and a unit which 
renders possible a relation of consciousness (of which the 
factors are consciousnesses) must be conscious. The source 
of the relation and of the consciousness of it are the 

21. * But this very transcendental unity of apperception 
forms a connection according to laws of all the possible phe- 
nomena which can ever appear simultaneously in experience.' 1 

The sequence of thought seems to be as follows : Self- 
consciousness on the part of the mind, of the identity of the 
function by means of which it connects the manifold syn- 
thetically in a cognition, renders possible the unity of con- 
sciousness in the sense of a connection with a given sensation 
of all related possibilities of sensation so as to form this or 
that object, of which we can have a single experience. It 
follows that the original and necessary (transcendental) 
consciousness of self is also consciousness of the one tran- 
scendental object, relation to which constitutes the ' objective 
reality' of our experience. Consciousness of the identity 
of its own function is also consciousness of unity in the 

1 P. 572; p. 201, Tr. 


correlation of all that appears to it, of all phEenomena (or, 
as Kant says, ' of the unity of synthesis of all phenomena 
according to concepts'; concepts really = conceived laws), and 
such unity is the 'transcendental object,' the conception of 
which compels us to interpret every experience as consistent 
with all the rest, and to reject as unreal every ' seeming 
experience,' i.e. every interpretation of feeling which con- 
flicts with the general system of experience. 1 

With Kant, then, the transcendental object and tran- 
scendental subject are the same. The presence of an eternal 
and unchangeable self to all phenomena at once makes 
them an order of nature, and makes our experience of them 
one connected system. ' Order of nature ' and c unity of 
experience ' are only two aspects of one and the same 
function of the eternal self, which we call object or sub- 
ject, according as we look on one or the other of these 

22. We have consciousness, then, of such object or 
subject (in Kant's language, we ' think ' it), but we have 
not ' knowledge ' of it, because it is not given in any intui- 
tion, and intuition is necessary to constitute knowledge. 
There is no phenomenon, and no sum of phenomena, of 
inner or outer sense of which we can say ' this is it ' or 
' these are it.' It renders possible experience as an experience 
of objects, but is not an object of which there can be 
experience. Is it, then, the ' thing in itself ? Yes, accord- 
ing to Kant, it is that 'thing in itself which rentiers 
possible ' nature in the formal sense.' It seems as if when 
he wrote the first edition of the Critique he was coming to 
regard this as the sole ' thing in itself,' but the final view, 
into which he had settled down when he wrote the Prolego- 
mena, was that there was another 'thing in itself,' which 
renders nature possible in the material sense, the cause of our 

There seems to result an opposition between the source 
of our experience, qua sensation, and the source of it, qua 
order of sensations. 2 

'In the following sentence, 'For of its function,' and only arises upon the 

this unity of consciousness would be view of this junction and the empirical 

impossible, &c.' does Kant mean to say synthesis which results from it ? 

that the ' thought of its own identity ' 2 ' How is nature at all possible in the 

on the part, of the mind is other than material sense, as to intuition, [I mean 

this ' self-consciousness of the identity nature] considered as the complex of 


But the whole drift of the i deduction of the categories,' 
as it appears in the first edition of the Critique, is to show 
that ' objects ' are laws of relation between phenomena, con- 
stituted by the synthetic self-consciousness which ' makes ' 
nature. In fact, when we set about accounting for a sen- 
sation, how do we do it ? By ascertaining uniform relations 
under which it occurs. These are the 'nature' which the 
understanding ' makes,' and which in turn makes our sen- 
sitive experience, so far as anything can be said to do so. 
No doubt they presuppose something else, but that is the 
eternal subject, not any 'substance' or 'thing in itself 
independent of and opposed to this. 

23. The great embarrassment throughout Kant arises 
from his view of ' phenomenon ' as something immediately 
given apart from its determination mediately, through con- 
ceived relations. So he says ' phenomena are the only 
objects which can be given us immediately, and that which 
in the phenomenon refers immediately to the object is called 
intuition.' l Phenomena are ' immediately given,' yet in the 
phenomenon a distinction has to be made between that 
which relates immediately to the object and that which does 
not. Let the phenomenon be this table. The consciousness 
so described would contain according to Kant an element 
of intuition and an element of conception. Its qualities 
consist in relations which we conceive, which, according to 
Kant, are not immediately given, not intuited. The intui- 
tion is represented by the this. It is the consciousness of 
something here and now affecting me. This in Kant's 
language relates immediately to the object. But what is 
the object? Only a relation, only a necessary sequence of 
certain sensations on others — a relation constituted by the 
unity of understanding, and necessary on account of the 
presence of one subject to all feelings. Such an object is 
essentially constituted by conception. Is there, then, no 
meaning in Kant's view that it can be ' intuited,' as dis- 
tinct from the transcendental object which ' we can no longer 
intuite ' ? 

phenomena? how are space, time, and totally distinct from those phenomena ' 

that which fills both — the object of (Prolegomena zu einer jeden Minftigen 

sensation — in general possible ? The Metaphysik, sec. 36 ; translated by 

answer is, By means of the constitution Mahaffy in Kant's Critical Philosophy 

of our sensibility, according to which it for English Headers, vol. iii. p. 99). 

is specifically affected by objects, which ' P. 573; p. 202, Tr. 
are in themselves unknown to it, and 


One may distinguish a perceived object, as that to con- 
stitute which there must be a sensation, though a sensation 
determined by conceived relations, from an object such as 
' nature/ which means the thought of a nexus between all 
possible sensations, which, therefore, cannot specially deter- 
mine any one sensation or group of sensations. For Kant's 
distinction between intuition and conception, therefore, it is 
better to substitute one between conception as determinant 
of particular feelings, and so constituting perception, and con- 
ception of laws of relation, as apart from the feelings which 
they determine. 

24. From Kant's way of putting the matter — as if there 
were sensibility giving ' phenomena ' or intuitions on the 
one side, and unity of understanding on the other — arises 
the question how the former should correspond to, come 
under laws prescribed by, the latter. Kant meets this very 
well when it is put as the question, how nature should e con- 
form to our subjective apperception,' by showing that with- 
out such apperception there is no nature. 1 But in the other 
form he fails to meet the question, because he is always 
speaking as if there were objects of intuition independently 
of * transcendental apperception,' or unity of understanding. 2 

No one, of course, can suppose that ' unity of understand- 
ing ' in the abstract = ' nature ' as the complex of phenomena. 
Such ' unity ' is nothing real apart from the multiplicity of 
phenomena, any more than these apart from it. What is here 
objected to is the notion that phenomena, as affections of 
sensibility, are due to the operation of unknown ' things in 
themselves,' as opposed to that unchangeable subject, which 
maybe called a 'thing in itself in distinction from phe- 
nomena, because, though conditioning and realised in phe- 
nomena, it is not any one or all of them. The cause of any 
phenomenon, on its ' material ' as well as on its ' formal ' 
side, as sensation no less than as conceived, lies in its rela- 
tion to all phenomena, in the system of nature, and this the 
unchangeable subject renders possible. 3 

25. The unity of understanding ' makes nature.' The 
unchangeable self in ' relation ' to the multiplicity of repre- 

1 P. 576; p. 206, Tr. also p. 583; p. 216, Tr. 'Empirica 

2 Cf. p. 574 ; p. 204, Tr. ■ But the laws, indeed,' &c. 

possibility,' &c. with p. 576; p. 206, Tr. 3 See Kant's own words, p. 577; 

' That nature should conform,' &c. Cf. pp. 207, 208, Tr. 


sentations constitutes this unity. Hereupon we are tempted 
to say, ■ Either this implies that nature is a creature of my 
own, and, if so, what becomes of its objectivity ? Or else the 
unchangeable self must be God. Yet what self do I know 
of but my own, which is merely the " generalised abstraction 
of my continuous feeling " ? At any rate the representations, 
the manifold consciousness, which this subject is supposed 
to determine, are mine ; and how can the subject which de- 
termines any representation (Vorstellungeri) be other than my 
self? So we again lose hold of objectivity, as that which, in- 
dependently of ourselves, determines our consciousness.' Kant 
himself does not distinctly meet these difficulties. In answer 
to the question, What self do I know of but my own? he 
would say, You know no self but the empirical, i.e. the suc- 
cession of phenomena of the inner sense, because to con- 
stitute a knowledge there must be intuitions corresponding 
to conception, and only to the conception of the empirical 
self are there intuitions that correspond. But you can think 
an unchangeable subject of the changes in consciousness 
which we call phenomena (phenomena of the outer no less 
than of the inner sense), and you cannot merely think it, but 
know that in virtue of the functions which it exercises there 
is such a subject, because otherwise the changes would not 
be changes, or (which is implied in calling them changes) 
connected in one experience ; otherwise, in short, there 
would be no ' cosmos of experience.' If asked, Is this un- 
changeable subject God ? Kant's answer, or the answer in 
his spirit, would be, In calling it God you are trying to know 
that which you cannot know, because no phenomena represent 
it. Under the term ' God ' you are mentally applying to it 
predicates which do not stand for any real knowledge ; you 
are trying to say what the unchangeable subject is, whereas 
you are only entitled to say that it is. The intuitions, or 
phenomena, which it connects, and which, as so connected, 
form one world, are not intuitions of it or appearances of it. 
We have no intuition of it, and therefore cannot know it. 
No object is given us which corresponds to it (for the tran- 
scendental object merely = the law of synthesis which it 
constitutes), but the moral life is an endeavour, for ever in- 
complete, to construct such an object. 

26. So far we keep within the limits of what Kant in 
effect says. Going beyond these we may remark that when 


I oppose myself and my consciousness to the objective world, 
and say that a ' nature,' constituted by the presence of the 
self to representations, has not the objectivity which we look 
for in nature as that which determines our experience, I am 
tacitly taking myself and my consciousness to be merely what 
Kant calls the ' empirical ego,' merely the succession of 
representations to inner sense. But this is a ' false abstrac- 
tion.' Inner sense has no reality apart from outer, nor have 
phenomena of outer and inner sense any reality apart from 
the unchangeable subject through relation to which they 
become one cosmos. ' Objective nature ' must indeed be 
something else than ourselves and our states of consciousness 
as we are apt to understand these when we falsely abstract 
our states of consciousness from their conditions and our- 
selves from relation to the world ; but it does not follow that 
it is other than our states of consciousness in their full 
reality, i.e. in the fulness of those relations which presuppose 
relation to an eternal subject. I do not e make nature ' in 
the sense that nature = a succession of states of consciousness, 
beginning with my birth and ending with my death. If so, 
the ' objectivity ' of nature would doubtless disappear; there 
would be as many ' natures ' as men. But only by a false 
abstraction do we talk of such a succession of states. Their 
reality lies in eternal relations ; relations which are there 
before what I call my 'birth,' and after my < death,' if 'before' 
and ' after ' had any proper application to them ; and only 
through these relations are they known ; only through them 
do they form an experience. That kind of subjectivity which 
alone is incompatible with their being objective, i.e. deter- 
mined by permanent and necessary laws, lies merely in our 
misunderstanding of them. ' But how,' it may be said, ' can I 
misunderstand them if I am the eternal subject out of rela- 
tion to which their reality, as an order of nature, arises ? ' 
The eternal subject is me as ego, but not as an ego determin- 
ing all phenomena. If it were not me, my knowledge would 
be impossible ; there would be no nature for me. If it were 
me in its full reality, as the subject determining all phae- 
nomena, my knowledge would be all knowledge. 

27. 1 We have spoken of Kant's 'deduction of the pure 

1 [See above, Section 9. The present section is from a different MS., but seems 
to belong both in time of composition and subject to the preceding sections, the 
substance of which it repeats in a shorter and more general form.] 


conceptions of the understanding ' as explaining how there 
come to be certain relations objectively necessary, in the sense 
that they are not relations by which we may or may not 
connect phenomena according to our particular habits 
and tendencies, nor even relations by which it has become 
instinctive and unavoidable to mankind to connect them, 
but relations without which there could be no connected ex- 
perience, or (which to him is the same) no experience of 
objects at all. A sensitive experience is not an experience 
of objects unless the past feelings which any present feeling 
recalls are connected with it as appearances of the same 
thing. No doubt the notion of there being a multitude of 
separate things or substances in nature is one that has to be 
abandoned, but only in order to give place to the conception 
of nature as a uniform system. A more comprehensive 
conception of identity is substituted for a narrower one; but 
the judgment that all varieties of feeling represent a change 
of what remains the same, is from first to last the condition 
without which our feelings would not be experience of an 
objective world. 

It might seem that the natural course for Kant to take 
would have been to trace this judgment of identity to what, in 
fact, he believed to be its source — viz. the equal presence to all 
feeling of a thinking subject — and then to exhibit its various 
forms in a classification of the relations by which we connect 
phenomena. On the contrary, the first impression at any 
rate of his procedure is that he takes 'pure conceptions' 
(i.e. conceptions not derived from association of feelings) for 
granted, and then, in the 'deduction,' deals with the question 
' how these can relate a priori to objects ' without really 
showing that there are such conceptions. 'Once take the 
true view,' it may be said, ' that these conceptions are merely 
derived from the experience of objects by abstraction and 
generalisation, and the difficulty about their relation to 
objects disappears.' But in fact Kant's ' deduction,' though he 
defines it, as above, in such a way as in appearance to take 
' pure conceptions ' for granted, really (especially as stated in 
the first edition) amounts to a refutation of the view just 
stated. The * deduction,' in his language, answers the 
aucestio juris in regard to pure conceptions (i.e. the question 
"of their objective validity), the qucestio facti being already 
settled ; but, in fact, it is to it, in particular to the section 

VOL. II. d 



about the syntheses of apprehension, imagination, and 
recognition (omitted in the second edition), rather than to 
the previous sections, where he treats the quwstio facti, that 
we have to go for proof (so far as he furnishes an) 7 ) that 
there are certain conceptions which, being necessary to the 
experience of objects, cannot be derived from them. 1 

1 It would seem as if on p. 99 (p. 63, 
Tr.) Kant was using ' categories ' as 
equivalent to conceptions of a kind to 
constitute which, as he afterwards finds, 
' schemata ' corresponding to categories 
are necessary. These are there repre- 
sented as resulting from a unification of 
the manifold in intuition. On the same 
page, however, on which he seems to 
speak of them in this way, he also 
speaks of ' pure conceptions which give 
unity to the synthesis of imagination,' — 
which do the unifying work from which 
the categories, in the previous sense, 
result. This latter way of speaking 
best corresponds to his after-use of the 
term « categories ' (p. 146 ; p. 113, Tr.). 
It is difficult, however, if we separate 
the unifying function of understanding 
from the schemata through which it 
constitutes definite conceptions of ob- 
jects, to see how it comes to be spoken 
of in the plural as ' pure conceptions ' 
(above) ; in other words, how, as a 
unifying function separate from the 
schemata, it differs from the simple 
unity of apperception. On p. 128 (p. 
94, Tr.) he speaks of categories as simply 
apperception in its application to the 
manifold of intuition in general. This 
intuition, again, as he explains on p. 
126 (p. 92, Tr.), must be sensuous. 
(Cf. p. 128 ; p. 89, 7K) 

It appears, then, that according to 
Kant we have (a) the unity of apper- 
ception, the function of understanding 
in general ; (b) the same in its possible 
application, but not actually applied, to 
the manifold of sensuous intuition in 
general. It is pluralised in virtue of 
the plurality of its possible applications. 
Hence it is spoken of, not as the func- 
tion of understanding, but as functions 
of judgment or conception, and even as 
pure conceptions, which yet, we are told, 
do not amount to actual conceptions be- 
cause they still await schemata in order 
to realise their application to objects. 
These are the categories as usually 
spoken of (though not as spoken of on 
p. 99), being not the results of the uni- 
fication of the manifold in intuition by 

understanding, but the possibility of 
such unification, (c) These ' functions,' 
as supplemented by, or acting through 
and upon the transcendental synthesis 
of imagination, and thus becoming con- 
ceptions of objects having a pure or 
transcendental content, — a content re- 
sulting from the unification of the 
manifold of pure intuition (or of the 
forms of intuition) by transcendental 
apperception. These are the concep- 
tions enumerated in the table of cate- 
gories. (The categories of substance 
and of quantity clearly mean more than 
the mere function of unification unap- 
plied.) Cf. p. 152 (p. 119, Tr.). The 
pure conception, as just described, re- 
sults from the combination of three 
factors there particularised, as distinct 
from the ' cognition of experience ' to 
which it is applied. (For the distinc- 
tion of 'pure a priori conceptions' in 
sense (<?) from the categories, see p. 
142 ; p. 108, Tr., « and finally,' &c.) 

What, then, is the 'schematism of 
the categories ' ? It is a process already 
involved in ^ (c), as mediating between 
the categories in sense (b) and pheno- 
mena, and rendering possible the appli- 
cation of the one to the other. 

On p. 126 (p. 92, Tr.) Kant seems to 
say that the category, minus schematism 
or the work of transcendental imagina- 
tion^ already involves synthesis of the 
manifold of intuition in general, as dis- 
tinct from synthesis of the manifold of 
sensuous intuition. Can this distinction 
be substituted for that between (b) and 
(c) above, as representing Kant's general 
view of the distinction between cate- 
gories and categories plus schemata ? 
The opening passage of § 20 {ibid.) is 
against this interpretation. On the 
whole, the distinction which generally 
fits best is that between categories, as 
functions of unity not yet applied to a 
manifold of intuition at all, — as pure 
conceptions Svvdfiei, — and pure concep- 
tions as applied through schematism 
to the manifold of pure intuition, and 
thus capable of application to empirical 


MATICAL principles: 

\Krit. d. r. V. pp. HO-165 ; pp. 107-132, Tr.] 

28. The peculiarity in Kant's view of the ' schemata,' as 
a tertium quid between the categories and sensible intuitions, 
arises from the separation which he makes between these as 
constituting severally the form and the matter of knowledge. 1 
On p. 123 (Tr. p. 89), having said that 'the manifold to 
be intuited must be given previously to the synthesis of the 
understanding, and independently of it,' he seems in the 
sequel to take this as equivalent to saying that under- 
standing can only operate on objects given to it (as opposed 
to by it), and so given in intuition. But is there not here a 
fallacy ? It does not follow because the 'manifold for intuition' 
is given independently of understanding, that intuited objects 
are so given. It may be the action of the understanding 
that converts the ' manifold for intuition ' into an intuited 
object; and Kant in some places seems to imply that this was 
his actual view, as it certainly is that to which his doctrine 
logically leads. 

However, admitting the separation between objects as 
given in intuition on the one side, and the categories on the 
other, Kant reflects that the categories, as apart from the 
'objects,' can yield no synthetical judgments. This may be 
readily admitted; the only question would be whether, as 
thus apart, they could yield any judgments at all. As he 
puts it, a synthetical judgment connects with a given con- 
ception something which is not thought in it, and conse- 
quently so connects it by a relation, which is not one of 
identity or contradiction (these being the relations between 

1 See p. 82, ' Ohne SinDlichkeit eine Erkenntniss u. s. w.' : Tr. pp. 46, 
u. s. w.' ; and cf. p. 93, 'Da keine 57,118. 
Yorstellung u. s. w.' ; and p. 151 , ' Wenn 

d 2 


a given conception and its content). What is the medium 
of this connection — the synthesis between a given concep- 
tion and an object — as (a) between the categories and objects 
of intuition, (b) between one conceived object and another? 
The answer is, the internal sense, as containing all our repre- 
sentations, and its a priori form, time. It is the determina- 
tion of the form of this sense that constitutes the medium 
between the categories as such and objects of intuition as 
such, in the shape of transcendental determinations of time 
or schemata; while it is internal sense, as determined by 
the categories through these schemata — in other words, our 
experience — that forms the medium of connection between 
one conceived object and another, and so renders possible 
synthetical judgments — judgments respecting matters of fact. 
29. In this doctrine two views are mixed up, one true, 
the other questionable. There is no doubt that only through 
sensuous experience, as determined by 'unity of apperception,' 
and only thus becoming a connected system at all, is a world 
given us to be known : only as relating to objects of such a 
world have conceptions any meaning. No ' instructive pro- 
positions ' are possible except as connecting conceptions in 
virtue of a common relation to a sensible object which they 
alike determine. There is no doubt, further, that sensuous 
experience, as determined by unity of apperception (in this 
perhaps there is a variation from Kant), involves the relation 
of time, so that all judgments about it either presuppose 
(as judgments respecting number and magnitude) or consist 
in (as judgments concerning events related in the way of 
cause and effect) some ' determination of time.' For reasons 
elsewhere given, 1 it is improper to speak of such experience 
as specially inner, according to the only meaning in which 
the opposition between outer and inner sense can be main- 
tained — i.e. as experience interpreted as a succession of 
changes in oneself as opposed to outward things. But, 
under this correction, there is so far nothing to find fault 
with in Kant's doctrine. His c principles of pure under- 
standing ' truly represent conditions of experience — judg- 
ments not derived from experience by generalisation (therefore 
miscalled highest generalisations), but implicitly contained 
in the most primary experience. Certain determinations of 
time being involved in the application of thought (unity of 

1 [See below, Section 57.] 


apperception) to sense, we may abstract these from the 
judgments representing the conditions of experience and 
call them ' schemata ' ; but it is a mistake to speak of these, 
which are abstracted from the concrete reality of sense 
determined by thought, as if they rendered possible the 
application of ' pure conceptions ' to sensible objects. Kant 
falls into this mistake from a false separation between the 
categories and objects of intuition, as if either had inde- 
pendent existence. If the categories and phenomena were 
really so ' heterogeneous ' that a tertium quid must be found 
to mediate between them, whence should this tertium quid 
be got? Kant leaves no alternatives but these — viz. that (a) 
it originates in thought, or (b) in sense, or (c) results from 
the determination of the latter by the former. If either (a) 
or (b) were a true account of it, the tertium quid would not 
serve the purpose of mediation ; if (c), the determination of 
sense by thought, of phenomena by categories, which it is 
said to render possible, is presupposed by it. 

Kant does not seem to admit that the e unity of apper- 
ception ' (and the categories, except that they are differen- 
tiated or pluralised through application to objects, are just 
this unity) is as necessary to the giving of objects as to that 
connection of them, when given, in one experience without 
which, according to his own language, there is no ' objective 
reality.' He writes as if the mere w manifold to be intuited,' 
the succession of feelings as it might be if there were 
nothing distinguishing itself from the succession to combine 
them, were already objects. We have, then, objects apart 
from which pure conceptions are empty (as it is admitted 
that conceptions are except in relation to objects or phe- 
nomena), and which conception (or* apperception) has no 
part or lot in — which are there quite independently of it. 
Thus arises the problem of finding something to mediate. 
In fact objects, in that sense in which it is true that thought 
is empty except in relation to them, are not a mere ' manifold 
to be intuited,' and are what they are as objects through a 
synthesis affected by apperception. 

30. The two unadjusted sides of Kant's doctrine appear 
together on p. 151 (Tr. p. 118). That which gives ' objective 
reality ' to cognition (strictly, gives such reality to concep- 
tions that they become cognition or knowledge) according 
to the first paragraph, itself depends, according to the 


second, on a synthesis according to conceptions. ' To give 
an object ' = ' to apply the representation of it to experience, 
actual or possible.' The mere succession of sensations does 
not amount to the giving of an object. The antecedent 
condition of this is the consciousness of something which 
implies determination of the successive feelings by a subject 
not in the succession. The consciousness of a qualified 
object having been arrived at, this remains with us as (in 
Kant's language) a 'representation,' the consciousness of 
which does not amount to the giving of an object. I have 
a representation of the table in my study, or of the colour of 
a geranium, or of the weight of the atmosphere ; but these 
are not yet real cognitions — objects corresponding to them 
are not given — unless the representations are 'related to 
actual experience ' in a perception, or to possible experience 
in the knowledge of the conditions under which such per- 
ception might be had. Perception, however, is not equiva- 
lent to sensation. When I perceive the colour of a certain 
geranium, a sensation or sequence of sensations is deter- 
mined by prior conceptions, which, through the circum- 
stances of the sensation, become connected with other con- 
ceptions in a synthetical judgment, e.g. ' this red geranium 
grows on ferruginous soil.' Thus that ' application of the 
representation of an object to experience,' through which 
the object is said to be given, is its connection with a system 
of facts, which only exists in virtue of the continuous refer- 
ence of all sensations to an object, ' the object of phenomena 
in general.' This reference, through which sensations be- 
come facts qualifying something, and thus mutually deter- 
mining qualities, is the work of ' unity of apperception.' It 
is a ' construction ' (which seems the best equivalent for 
Kant's ' synthesis ') ' of an object of phenomena as such ' 
(i.e. of an object of which all phenomena are considered 
appearances), 'according to conceptions,' i.e. determined by 
certain thought relations. Thus Kant says (loc. cit.) ' experience 
depends upon the synthetical unity of phenomena, that is, 
upon a synthesis according to conceptions of the object of 
phenomena in general, a synthesis without which experience 
never could become knowledge, but would be merely a 
rhapsody of perceptions, never fitting together into any 
connected text, according to rules of a thoroughly united 
(possible) consciousness, and therefore never subjected to 


the transcendental and necessary unity of apperception.* 
The c general rules of unity ' of which he speaks below are 
the ' axioms of intuition/ &c. expounded in the sequel. Cf . 
the passage on pp. 118-119 (Tr. p. 84). ' It is the unity of 
consciousness alone that constitutes the possibility of repre- 
sentations relating to an object, and therefore of their 
objective validity, and of their becoming cognitions, and 
consequently the possibility of the existence of the under- 
standing itself. ? It is not through sensation, then, that 
objects are given, but through experience, i.e. unity of 
apperception as exercised upon sense. And since the cate- 
gories themselves are nothing else than the forms of this 
unity, as so exercised, nothing is needed to mediate between 
them and objects. The ' Transcendental Analytic ' would 
have been much simpler if the account of the categories 
prior to the ' Deduction ' had been omitted. The categories 
then would not have appeared in that separate form in 
which they are made to correspond to the classification of 
logical judgments (a classification which is only of value in 
relation to the syllogism, and which represents as little as 
the syllogism the process by which intelligent experience is 
formed). We should have had (1) what is fancifully called 
the ' Deduction of the Categories,' exhibiting the unity of ap- 
perception, derived from the presence of the ' transcendental 
ego ' to all feelings, as the condition of the possibility of 
experience, and then (2), without surplusage of distinction 
between ' categories ' and ' schemata,' an account of the 
6 principles of pure understanding ' (as given in the third 
section of the ' System of Principles ') i.e. of ' the general 
rules of unity in the synthesis of phenomena,' as arising 
out of the application of the thinking unit to the ' manifold 
of sense,' and thus involving ' determination of time.' 

31. By means of number — the schema of quantitas as 
determination of time — we are able to know phsenomena as 
quanta (this is what from Kant's account of number, as the 
schema of quantitas, we should expect him to show). The 
account of number is given on p. 144 (Tr. p. 110). A ' homo- 
geneous intuition ' seems to mean an intuited object as made 
up of homogeneous parts. Kant's language is naturally 
taken to mean that such an intuition is given, and that we 
proceed to apprehend it, to take in its several parts, and 
that number results. But in fact such an object is only so 


called, or said to be made up of parts, proleptically. It comes 
to be ' a homogeneous intuition ' in this sense through the 
' unity of the synthesis of a manifold,' the manifold being 
successive feelings which through this unity (through being 
held together in relation to one subject) become homo- 
geneous parts of one intuition. What Kant himself meant 
I am not sure. If he meant that, given a feeling fixed as 
an object to be attended to (an intuition, in this sense), then 
in that process of attending to the felt object through which 
it is apprehended as a collection or whole of homogeneous 
parts (an extensive quantum), time is generated as the succes- 
sion between the acts of attention, and number results from 
the synthesis of the units of time thus generated, then he is 
intelligible. 1 Number is thus, according to Kant, quantity 
of time. But anything else can be counted as well as time ; 
e.g. as Kant shows under the * Axioms of Intuition,' parts of 
space. We may admit that time is generated in the process 
of attention to an object of intuition (as understood above), 
and that number results from the unity in the synthesis of 
the data of successive "acts of attention; but it does not 
follow that these data in their most primitive form are 
times, or that what is first counted is time. What is 
generated as above is time as a relation of succession, not time 
as a numerable quantum. It would seem as if Kant misin- 
terpreted the truth that number is the result of counting, 
and that counting as a process implies succession or time, into 
the notion that what is primarily counted is moments of time. 
Number, then, is in no special sense — as it is according 
to Kant's account of it as a ' schema ' — a determination of 
time. It is a result of counting, and any object that is 
apprehended as one through successive acts of attention is 
so far — whatever else it may be — a numerable object, an 
object made up of homogeneous parts that can be counted, 
that has number. There is no ground for the distinction 
between quantitas, as the category or pure conception, 
number as the schema, and the quantum as the object of 
intuition to which, by means of the schema, the category 
is applicable. Quantity is the possibility of being counted, 

1 ' Why,' it may be asked, ■ should it Kant, time is the relation of succession 

be said that time is generated in the between states of consciousness. In' 

process of attention ? ' A certain time giving rise to a succession of states, 

is occupied by it, do doubt, but that is then, we ipso facto generate time, 
quite a different thing. According to 


the relation of addibility or numerability between objects. 
Number is the actuality or result of being counted. Quanta 
are objects, between the parts of which the relation called 
quantity subsists, of which the parts can be counted. The 
condition of there being such a process as counting, or 
objects (quanta) to which it is applicable, is the interaction 
of thought and sense. In order to there being number 
there must be present to successive different feelings a 
subject distinguishing itself from them, which can retain 
their mere differences and at the same time put them 
together as one to one. In order to there being numerable 
objects (quanta) there must have taken place (a) a dis- 
tinction by a thinking subject of a feeling from itself, and 
the presentation of it to itself as an object to be attended 
to, and (6) a reference to this object of the data of succes- 
sive acts of attention as its parts, 

32. Phenomena^ sensible objects. To constitute a sen- 
sible object, there must be (a) sensation ; (b) the form of 
intuition, i.e. distinctness between now and then, here and 
there, of what is given in sense ; (c) synthesis of apprehen- 
sion, by which the sensible data, thus manifold and distinct, 
are combined in one object. The constitution of a phe- 
nomenon, then, or intuited object (and this is the sense in 
which Kant uses 'intuition' in the 'Analytic of Principles'), 
involving synthesis of manifold heres and theres, nows and 
thens, which are homogeneous with each other (however 
different what is perceived here may be from what is perceived 
there, &c), is ipso facto 'generation' of a determinate space 
or time, as made up of spaces and times, i.e. of a ' quantum ' 
as a whole of homogeneous parts. It is not that a deter- 
minate space or time is first given and that then it is broken 
into parts. It is that in the process of apprehending the 
manifold of an intuition, when one ' here ' has taken the 
place of another which accordingly becomes ' there,' I retain 
the consciousness of the ' there ' and add it to the * here,' 
and so again with both when I come to the next ' here,' 
thus 'rendering possible the representation of the whole 
through the representation of the parts,' and so consti- 
tuting, according to Kant's definition, an extensive quantity. 
Mutatis mutandis, the same applies to time. 

Thus every pure intuition (or ' formal intuition '),' in 

1 P. 132, note ; Tr. p. 98. 


virtue of the synthesis of a manifold in space or time which 
is necessary to constitute it, is an extensive quantity ; and 
'empirical intuition is possible only through pure intuition.' 1 
Empirical intuition is of this or that sensible object, as 
occupying space or time. As thus occupying space or time, 
the synthesis of apprehension through which it is given 
involves that which constitutes determinate spaces or times 
(though much more than this, viz. the synthesis of sensible 
data given in those spaces or times), and thus the empirical 
intuition, though much else, is an extensive quantity, and 
all the ' mathematics of extension ' are applicable to it. 
Hence this ' transcendental principle ' (viz. all ' intuitions 
are extensive quantities') ' enlarges our a priori cognition.' 
' The synthesis of spaces and times as the essential form of 
all intuition is that which renders possible the apprehension 
of a phenomenon, and therefore every external experience, 
consequently all cognition of the objects of experience ; and 
whatever mathematics in its pure use proves of the former 
must necessarily hold good of the latter.' 2 

In all this it does not appear that the schema number, 
as quantitas of time, renders possible the knowledge of 
phenomena as extensive quantities. The determination of 
spaces and times alike, as Kant here describes it — the 
process by which we come to intuite so much space or time 
— is a process of ' adding one to one ' in which number is 
c generated.' Just as Kant says that ' empirical intuition is 
possible only through pure ' (since the former involves the 
latter), so one may say that the synthesis of a manifold in 
space and time (which yields pure formal intuition) is possi- 
ble only through a synthesis (adding of one to one) of the 
manifold, as such, which = counting, and yields number; in 
other words, that the mere adding of one to one is the pre- 
supposition of the particular adding of a here to a there, a 
now to a then. 

33. It does not make any real difference whether you say 
that the synthesis of pure intuition is the presupposition 
of, or an abstraction from, that of empirical intuition ; nor, 
in like manner, whether you say that number is the presup- 
position of, or the abstraction from, extensive quanta ; so 
long as no opposition is intended between abstraction as the 
work of thought and the reality as something with the 
constitution of which thought has nothing to do. If it is 

1 P. 168; Tr. p. 125. 2 Ibid. 


said that space and time are ' mere abstractions,' this is apt 
to mean that there are real objects in space and time given 
to thought, without any co-operation on its part, and that it 
proceeds to abstract pure space and time from them. This 
is a delusion, but there is none in regarding mere space and 
time as abstractions made by thought from its own concrete 
construction. So it is a mistake to say that number is 
an abstraction made by thought from real quanta given 
independently of it. The process which generates it is a 
process of thought involved in that which generates deter- 
minate spaces and times, as this again in the process of 
thought which generates concrete objects in space and 
time ; but the most abstract form of the process is that in 
which it generates mere number. 

34. We have seen how perceived objects (empirical 
intuition), in virtue of their form as intuited, i.e. as occupy- 
ing space or time, are extensive quanta. The reality of such 
objects, however, does not consist in the space and time 
which they occupy, but in their sensible qualities, i.e. in 
affections on our part referred to some external object. 
Such sensations, so referred or transformed into qualities, 
Kant calls * the real in phsenomena.' This ' real,' not 
occupying so much space or time, cannot be an extensive 
quantum, yet we estimate it numerically. An object affects 
us, is hot or heavy or bright or noisy, to such or such a 
degree of intensity; i.e. according to Kant, between its 
reality (the sensation on our part which we refer to it) and 
nothing (entire absence of sensation) there is a possibility of 
sensations, ' the difference of which from each other is alwaj^s 
smaller than that between the given sensations and zero, or 
complete negation.' * We do not need to traverse these in 
apprehension (as we do the parts of a space) in order to the 
apprehension of the real in phenomena, since 'apprehen- 
sion takes place by means of mere sensation in one instant, 
and not by the successive synthesis of many sensations, and 
therefore does not progress from parts to the whole ; ' 2 and 
therefore the real is not an extensive quantum ; but it has 
intensive quantity or degree as = ' that quantity which is 
apprehended only as unity, aud in which plurality can be 
represented only by approximation to negation.' 3 

Thus, though nothing can be known of the quality of a 

1 P. 160; Tr. p. 127. 2 Ibid. 8 Ibid. 


sensation a priori (i.e. before it is felt), we know a priori 
that whatever the sensation, as real or representing the 
real it must have a degree, because to know it as real is to 
know it as determined by opposition to absence of sensation, 
and we cannot so know it without representing to ourselves 
a continuous process by which the ' empirical consciousness ' 
rises from negation as = the absence of the sensation to 
reality as = its presence. As Kant puts it, when we say that 
a sensation is the real in a phenomenon, ' the real' means 
' the synthesis in an empirical consciousness,' or i synthesis 
of homogeneous ascension from up to the given empirical 
consciousness.' l 

35. According to this account of the matter, for a merely 
sensitive consciousness any given sensation would not be 
real, and would not have degree. It is at once real and has 
degree in virtue of an intellectual synthesis by which a 
transition of empirical consciousness from absence of feeling 
to the given one through infinitely reducible intermediate 
stages is presented to thought. 

The sensation, as such, is apprehended in a single 
instant. In this it is unlike a space, as made up of parts, 
which can only be apprehended in a succession of moments. 
As sensation, then, it is one, but the apprehension of it as 
real implies the conception of a process ' in which the 
empirical consciousness can within a certain time rise from 
nothing up to its given amount.' Thus ' plurality ' is repre- 
sented in it ' by ascent from negation,' or, which is the same, 
* by approximation to negation.' 

Bearing this in mind, we may understand how it is that 
Kant falls into the (at least) verbal contradiction of saying 
in one place that 6 quality belongs to sensation by means of 
its apprehension, in which empirical consciousness can within 
a certain time rise from nothings up to its given amount,' 
and in another that ' the real in a phenomenon has always a 
quantity, which, however, is not discoverable in apprehen- 
sion.' 2 The explanation is that no judgment of quality is 
involved in apprehension as = consciousness of sensation, but 
that such judgment is involved in the apprehension of it 
as the real in a phenomenon, because this involves the pre- 
sentation of the process through which ' empirical conscious- 
ness can within a certain time rise from nothing =0 up to 
its given amount.' In the first passage the words in a 

1 Pp. 164-165; Tr. p. 132. 2 Pp. 159-160; 7V. pp. 126-127. 


parenthesis, to make sense, should run, ' and that in virtue 
of a conceived process of apprehension, in which,' &c. That 
is to say, the apprehension here spoken of is not that spoken 
of as ( apprehension by means of sensation alone' (p. 127, 
Tr. ), but the ' synthesis of homogeneous ascension from 
up to the given empirical consciousness' (p. 132 Tr.). 

36. In the account of the ' anticipations of perception ' 
the characteristic of the real, as filling time, which is insisted 
on in the account of the schemata, does not appear. We should 
expect to find it shown that the category of reality (' reality 
in the pure conception of the understanding ') becomes 
applicable to sensation (so that sensation becomes the 
representative of the real) through a schema consisting in 
a determination of time. This schema is described as 
' Zeiterfullung.' It is difficult to see how a filling of time 
can result from the determination by understanding of time 
as the mere form of intuition — which is what a ' schema ' 
according to Kant is — and thus how a filling of time should 
be a schema at all. What is shown is that the knowledge 
of sensation as real, or as that to which the real corresponds, 
implies the representation of a possible transition from it to 
negation or from negation to it through infinitely divisible 
stages, in virtue of which it has degree. The representation 
of such a transition implies that of time as a relation of suc- 
cession, and of a different filling of each moment of the time 
in which the transition takes place. Each such filling in 
turn must in being known be represented as reached through 
a like transition from negation through the succession of 
differently filled times. It is true, then, that ' a determina- 
tion of the inner sense according to conditions of its form, 
viz. time,' l is implied in the knowledge of sensation as real. 
To know it as real is to know it as more or less intense ; to 
know it as more or less intense implies the conception of a 
process of empirical consciousness (or of a determination of 
the inner sense) by which it might descend to zero or have 
ascended from zero, and this is a conception of a process in 
time. Just so far, however, as this is a conception of a 
process in which moments are successively filled, it is not a 
conception of pure time, and does not correspond to Kant's 
account of the schema as • transcendental determination of 
time,' rendering possible the application of the category to 

1 P. 143; TV. p. 110. 



[See especially Krit. d. r. V. pp. 165-169; Tr. pp. 132-136.] 

37. The ' analogies of experience ' represent the modes 
in which perceptions mnst be connected if they are to form 
one experience, or (which is the same) to represent one 
world, or to become a cognition of objects (p. 165 ; Tr. p. 133) ; 
they are merely regulative, not constitutive principles, i.e. they 
do not give or enable us to construct intuitions, nor do they 
relate to the presentation of objects (phenomena) ; but, given 
the phenomena, they determine the way in which they must 
exist in relation to each other. Thus they are distinguished 
from ' axioms of intuition ' and ' anticipations of percep- 
tion, 5 inasmuch as these determine what every phenomenon 
must be, viz. an extensive quantity in respect of its relation 
to intuition, intensive in respect of its relation to sense. 
The ' analogies ' on the other hand do not enable us to say 
a priori what any phenomenon must be (thus ' they do not 
concern phenomena ') ; we are not able by means of them 
to anticipate ' in what respect the empirical intuition of it 
would be distinguishable from that of others ' (p. 167 ; Tr. 
p. 134), but only that, whatever the phenomenon, its existence 
must be determined in a certain way by relation to other 
phenomena, as forming along with them a series of changes 
determined by something unchangeable, or as related to them 
in the way of cause and effect, or in that of reciprocal action. 
The term ' analogy ' is borrowed from its use in mathematics. 
In mathematics it means an equality of proportions, such 
that, given two numbers bearing a certain proportion to 
each other, and a third, known to be related to another in 
the same proportion as the former two, we can tell what 
this fourth number must be. In philosophy it means a ru]e 
in virtue of which, given a phenomenon or perceived object, 
we can assert the necessary existence of another object 


related to it in a certain way, though we do not perceive 
this other object, and cannot (merely in virtue of the rule 
and without further experience) say what it is. But such 
a rule, though it does not by itself enable you to say what 
the other object is, puts you on the track for finding it ; it 
sets you, e.g., on the faith that there is uniform sequence 
of phenomena in time, to look behind the apparently shifting 
order of our sensations with the purpose of finding what the 
event really is which precedes a certain other event, and 
which, preceding it once, precedes it always. 1 

38. Kant's distinction between the knowledge represented 
by the ' axioms ' and ' anticipations ' as constitutive, in oppo- 
sition to that represented by the ( analogies ' as regulative, 
seems to involve two points. You can know a priori (a) that 
a phenomenon must have quantity, extensive and intensive, 
whereas of the phenomenon inferred in virtue of an analogy 
of experience you only know that it must exist in a certain 
relation to the given phenomenon, but nothing of its nature. 
We naturally ask whether the latter knowledge does not 
1 concern phenomena ' just as much as the former. Kant, 
however, seems to regard the phenomenon as having an 
intrinsic nature, in respect of its being intuited and perceived, 
distinguishable from the relations to other phenomena 
under which alone it is known as existing. Hence he speaks 
of the rules which determine the latter as not * concerning 
phenomena.' This distinction is very questionable. To 
know a priori that a phenomenon must have a quantity is 
merely to know that it must stand in a certain relation to 
other phenomena, and just the same is to know that it must 
have a cause. ' But ' (it may be said) ' the knowledge that it 
has extensive quantity is necessarily incidental to the appre- 
hension of it as an intuition, and that it has intensive 
quantity to the apprehension of it as real. It is thus inci- 
dental to the giving of the phenomenon as an object. The 
knowledge of the relations under which it necessarily exists 
is not so.' If it were possible to know an intuition as an 
extensive quantity, or a sensation as having degree, without 

1 P. 167; Tr. p. 135, 'Thus, if a in philosophy as distinct from quanti- 

perception is given us . . . discovering tative analogy. K.Fischer's explana- 

it.' In the sentence ' In this case,' &c. tion of the sense in which Kant speaks 

Kant is not thinking of the three con- of analogies of experience (Translation 

ditions of experience, which he after- by Mahaffy, p. 106) seems to me 

wards gives under the name ' analogies,' wrong. 
but of ordinary reasoning from analogy 


anything except intuition and sensation, there might be a 
valid distinction between such knowledge and that of an 
object as existing under a relation to another object, of 
which there is no intuition or sensation. But in the process 
of c representing a whole through representation of its parts,' 
one must conceive that part which is at any moment intuited 
as determined by the other parts, which have been, but are 
not being, intuited ; and in order to know a sensation as 
having degree one must conceive, without feeling, a series 
of possible sensations through which it would have to pass 
in order to become = 0, which means that the sensation is 
determined by a process — by a series of objects — neither felt 
nor intuited. What essential difference is there between 
such c cognition ' and that of a phenomenon as having its 
existence determined by another object, not present to sense 
or intuition, in other words, by a cause ? 

39. (b) The distinction means further that in virtue of 
( the axioms ' and ' anticipations ' you can know a priori 
what the particular quantity, extensive or intensive, of 
phenomena must be. Having learnt that the intensity of 
sunlight is 200,000 times that of moonlight, you do not 
need to await the sensation of sunlight to know what its 
degree must be. But how does this knowledge differ from 
the knowledge that, given a certain phenomenon, another 
event — its ascertained cause — though unperceived, must 
have happened? It is quite true that without experience 
you could not ascertain what the cause of the phenomenon 
is, but equally without experience you could not ascertain 
the relative intensity of sunlight as compared with moon- 
light. When on the faith of the uniformity of nature — or 
the * principles of the unity of experience ' — you have 
ascertained the law of relation, in one case as in the other 
your knowledge anticipates actual sensation. 

40. The ' principles of the mathematical use of the cate- 
gories ' are ' apodeictic ' — can have their truth demonstrated 
— because they represent conditions of intuition, and can be 
exhibited in intuition. 1 You can see that an intuition must 
have extensive quantity by simply attending to the intuition 
without considering anything else than the given intuition. 
The necessity is not contingent upon there being anything 
else than the intuition. On the contrary, the necessity of 

1 P. 154 ; p. 121, Tr. 


the principles of discursive cognition is dependent on objects 
of empirical intuition being given. These given, the dyna- 
mical principle of the understanding (e.g. the principle of 
substance) is necessary as the condition of their forming 
one experience. The necessity of the principle is contingent 
upon the phenomena being given to which it relates, and 
these phenomena are not given together in intuition. Given 
an event, e.g. its conditions are only known * discursively. 9 
They mast be objects of possible intuition, but in intuition 
they could only be presented successively — one ceasing to be 
so as the other comes to be so — and their relation to each 
other and to the event which they determine is not intuited 
at all. Knowledge of them, therefore, is ' through concep- 
tions, not intuitions ; discursive, not intuitive.' l 

It is difficult to see how the apprehension that an in- 
tuition has extensive quantity can be other than * discursive.' 
' The representation of the parts which renders possible the 
representation of the whole ' is a process of which the whole 
cannot be intuited at once. No relation can be intuited, in 
Kant's sense, because no relation — not even a relation of 
space or time — is in space or time. All knowledge, then, 
as of relations, is discursive, and its fault lies, not in this, — as 
Kant sometimes seems to think, — but in the fact that its 
range of discourse is so narrow. 

1 See p. 93 ; p. 57, Tr. ; and cf. p. 490 ; p. 447, Tr„ ' A priori conceptions,' &c, 

VOL. 11. 


f. the proofs of the 'analogies 
of experience: 

41. Expeeience is not a mere succession of perceptions, 
but the determination of an object by means of perceptions. 
As ( cognition of objects, 9 it implies that our sensations are 
referred, as signs or effects, to objects which do not pass with 
them — do not merely exist while sensation is felt — but are 
mutually qualifying elements of a permanent world. As 
Kant puts it, it implies a synthesis (a form of unity) not 
contained in perception, i.e. which does not arise out of the 
nature of perception, but which implies a unification of the 
manifold of perception, i.e. of sensuous affection, of such a 
kind as gives it a new character (synthetic), a character 
which does not belong to it merely as a manifold of sensuous 
affection, or as it is for a merely feeling subject. Every 
perception is determined a priori (before it actually occurs) 
by the necessity of being held together in one world with all 
other experience ; in other words, by the necessity of being 
referred to an object supposed always there. This reference to 
an object may take the following forms. Manifold percep- 
tions may be treated (a) as changing appearances of one 
thing, as ' sensible qualities ' of a substance, of which, as it 
is found that things, which we at first regard as independent 
of each other and absolutely permanent, are dependent and 
only relatively permanent, the conception expands into that 
of one constant sum of matter, (b) As a series of events of 
which each is so conditioned by that which precedes, that 
otherwise (without the antecedent) it could not have hap- 
pened. (In this case, the object always there is the uniform 
rule of sequence. The events pass, but the rule, that if one 
happens the others must, does not pass.) (c) As successive 
appearances of a system of things which coexist and are what 
they are in virtue of that coexistence. These three ways of 
determining perceptions a priori with reference to an objective 


world, which does not pass with them, Kant takes to corre- 
spond to the i three modi of time.' They are ' connections of 
objects in time in general,' ' relations of the existence of the 
manifold as it is objectively in time.' l 

42. * Time in general ' is opposed to the particular times 
at which perceptions happen to occur, in which * no character 
of necessity appears.' The order according to which the 
' manifold exists objectively in time ' is opposed to the order 
in which ' it is put together in time ' : e.g. it is a ' mere 
chance ' whether I see flame before I feel heat, or vice versa. 
If the relation of cause and effect belonged merely to the 
manifold ' as put together in time ' — to the order of our sub- 
jective apprehension — flame might be regarded indifferently 
as the cause or the effect of heat ; in truth, it is a relation 
of existence as it is objectively in time. The relation, Kant 
would say, between two phenomena as cause and effect is 
a relation of time (one must occur at a time preceding the 
other), but not of the times in which we may happen to per- 
ceive them. Often an antecedent event is not perceived. 
The shock of an earthquake is felt, but no one perceives the 
antecedent commotion in the bowels of the earth; the souud 
of a bell is heard before the motion of the clapper is seen, &c. 
It is a relation, then, of 6 time in general,' not of the times 
in which appearances occur to us ; and as ' time in general ' 
cannot be perceived, it is a relation arising, not out of per- 
ceptions, but out of ' a priori connecting conceptions.' The 
source of these (the ' analogies of experience ') is the equal 
presence to all feelings of the one thinking self, or, as Kant 
puts it, ' the general principle of all three analogies rests on 
the necessary unity of apperception in relation to all possible 
empirical consciousness at every time. 9 Inasmuch as the 
' original apperception relates to our internal sense,' or 3 more 
precisely, to the ' form ' of that sense, i.e. to distinctness of 
feelings in time, the unification of sensitive experience, 
which that ' apperception ' effects, must be a unification of 
it according to relations of time. Hence the rule which 
arises from it is, ' all empirical determinations of time must be 
subject to rules of the general determination of time ' (where 
6 empirical determination of time ' = all experience of objects in 
time), i.e. are subject to those rules according to which alone 
the succession of phenomena in time can form one world. 2 

1 Pp. 165-6 ; p. 133, Tr. 2 P. 166 ; pp. 133-4, Tr. 

k 2 


43. c Without the permanent, no relation in time is 
possible.' The succession of a upon b means that b is over 
before a begins. The relation between them (called succes- 
sion) cannot exist for either a or b, bat only for something 
present to each of them. If there were nothing but succes- 
sive events in the world (nothing but b vanishing before a 
begins, a before c begins, and so on), there could be no suc- 
cession. The possibility of any succession implies a * relative 
permanence,' and the possibility of everything being merely 
relatively permanent implies an absolute permanence. The 
same holds of change, which is a particular sort of succession 
— succession of different qualities. Again, 'it is only by 
means of the permanent that existence in different parts of 
the successive series of time receives a quantity, which we 
entitle duration. For in mere succession existence is per- 
petually vanishing and recommencing, and therefore never 
has the least quantity.' There must therefore be something 
not in succession but permanent that can carry on each 
vanishing moment of the succession and add it to the next 
in order to constitute quantity of time or duration. So far, 
good. £ But,' says Kant, * permanence is just another expres- 
sion for time, as the abiding correlate of all existence of 
phenomena, and of all change, and of all coexistence.' 
Time is such an ' abiding correlate ' because ' all phenomena 
«xist in time, wherein alone as a substratum, i.e. as the 
permanent form of the internal intuition, coexistence and 
succession can be represented.' Yet, on the other hand, 
i the permanent is the substratum of our empirical representa- 
tion of time itself, in which (sc. substratum) alone all deter- 
mination of time is possible.' ' Empirical representation of 
time' seems to be opposed to 'time in general,' which 
' cannot be an object of perception,' and to represent which, 
accordingly, a ' substratum' (matter) must be found in ob- 
jects of perception. 1 

I cannot see what meaning * time ' has except as a 
relation of succession, a relation of which the possibility 
supposes something other than the terms of the relation, 
something not in succession. It is a mistake to convert this 
relation into that permanent something, which is the con- 
dition of its possibility. It is a further mistake to speak, as 
Kant does, of ' permanence and coexistence ' as, along with 
' succession ' ' modi of time.' 

1 Pp. 169-170; pp. 136-137, Tr. 


44. When Kant calls time * the permanent form of internal 
intuition,' there is sense in this, so far as it means that 
time as =s succession is a permanent relation ; but then it is 
a permanent relation between the non-permanent or tran- 
sitory, just as transitory; the relation, namely, of inner 
intuitions to each other so far as each is over before the next 
begins. The relation time is not itself in time, the relation 
of succession is not successive. But it is one thing to say 
that succession (or time) is a permanent relation, and 
another to say succession = permanence, or that ' permanence 
is another expression for time.' ' Permanence ' expresses 
a relation to the transitory, or throughout the transitory, 
of that which is not so. Time is just the opposite of this, 
a relation of the transitory to the transitory. Thus to say 
that ' permanence is another expression for time ' is, strictly 
taken, nonsense. As permanence means a relation, so in 
this proposition time must mean a relation, and time as 
relation can only mean succession, which is the opposite of 
the relation of permanence. To say, however, that time is 
the permanent is a different matter, for in this proposition, 
as ' the permanent ' does not express a relation but some- 
thing related, so ' time ' may be taken for something other 
than a relation. No doubt, when Kant says that 'perma- 
nence is another expression for time,' he means ' time is the 
permanent.' He is thinking of time, not as a relation of 
succession, but as something in which all relations of time 
exist. 1 The question, then, is, whether you can properly 
speak of time as that which exists throughout all times, 
which is what the permanent means. 

Time is either the relation of succession, or a name 
representing something which we try to arrive at by adding 
times, ' times ' being what remains of the events between 
which the relation of succession subsists after the abstrac- 
tion of all determination of those events except such as arise 
from this relation. But the result of adding times is for 
ever incomplete, the addition being a process ad infinitum. 
There is really, then, no such thing as a time containing 
all times. If there were, it would not be the permanent, 
for it would be constituted by the addition of so many 

1 Cf. p. 170; p. 137, Tr. ' simul- 'there is only one time in which all 
taneity and succession are the only different times must be placed.' 
relations in time ; ' p. 173 ; p. 141, Tr., 


negations of permanence. Time, as=relation of succession, 
is permanent like all relations, in the sense of not being in 
time. A relation between events cannot be itself an event. 
But there is no reason for calling this relation the permanent. 
Strictly it is a result of the permanent, i.e. of the ' order of 
nature,' which, according to Kant's view, means the unifica- 
tion of the manifold through its relation to one subject. 

45. Time being either the relation of succession or dura- 
tion, succession is not a modus of time, but just time, and it 
is absurd to call either permanence or coexistence a modus of 
succession. Coexistence or simultaneity, as little as per- 
manence, can be spoken of as a modus of time. It is not 
properly times that are simultaneous. The only possible 
relation between one time and another is that of succession. 
Simultaneous events do not exist in different times simul- 
taneous with each other, but in one time successive on another 
in which other events coexist. Simultaneity is thus not a rela- 
tion between different times, but between different events in 
one time. It may be said, Why not reckon succession, coexist- 
ence, and permanence as c modi of time,' as each constituted 
by a different sort of i synthesis of different times ' ? But, 
at any rate, they are not co-ordinate. Succession = time. 
It is constituted by a synthesis of feelings of which one is 
passing as the other begins, but the synthesis must be per- 
formed in order that time may exist. Times being thus 
given, a further synthesis of their diversity may yield 
severally permanence and coexistence. But clearly it is a 
mistake to speak of ( the permanent' as = c time in general,' 
when it is only constituted by a synthesis in which times lose 
their character as times, which is that one is over before the 
next begins. 

46. * Substances (in the world of phenomena) are the 
substratum of all determinations of time. The beginning 
of some and the ceasing to be of other substances would 
utterly do away with the only condition of the empirical 
unity of time; ' l i.e. the condition under which alone the mere 
manifold of distinctions between beginnings and endings 
of feelings becomes one succession (time), is that something 
be equally present to them all. Why not say at once that 
this \ something ' is the eternal, that the eternal is thus the 
condition of there being time ? Kant would say, ' Because 

1 P. 173; p. 141, TV. 


the eternal is not an object of knowledge. If there were a 
possible intuition or phenomenon corresponding to the ' pure 
ego,' there would be a knowable eternal ; but it is not so. 
And the ' something ' which conditions the ' empirical unity 
of time ' — in relation to which determinations of time are 
possible — must be an object of knowledge, substance as 
phenomenon.' But is there any phcenomenal substance that 
does not begin and cease ? ' Not (according to Kant) any 
particular substance, but the matter of all.' But is matter 
in this sense a phenomenon? Not, if phenomenon means 
(as with Kant it does) that which can be perceived. All that 
can be perceived is some modification of matter which is not 
permanent. The matter or c something ' which really does 
not begin or cease, is the thinking self, as an object to itself, 
which is not in time at all, but is the condition of the possi- 
bility of time, and is only called permanent by a kind of 
metaphor and at the cost of contradiction. * Permanent ' 
(according to Kant) is that which is in all time ; but (a) 
time is not a possible all, and (b) the eternal is not in it. 

47. The point of Kant's 'proof of the second analogy' 
is that ' the subjective sequence of apprehension ■ must 
be ' deduced from the objective sequence of phenomena.' 
Against the doctrine that the relation of cause and effect is 
' nothing but uniform unconditional sequence ' he would 
have nothing to say. His point is that the existence of 
such a relation implies the determination of the sequence of 
our feelings by an ' order of nature ' other' than they, which 
= the unity of understanding, and is the a priori condition 
of the succession of feelings becoming an experience of an 
objective world. With Hume the ' subjective sequence of 
apprehension ' is everything. Connection between cause and 
effect is a determination of imagination to pass from an 
impression to the idea of its usual attendant. Feeling a 
has, as a matter of fact, so many times followed feeling b. 
This amounts to no order of nature. What we call so, 
according to Hume, is an expectation resulting from habitual 

Kant says, if the connection of cause and effect between 
a and b were merely a connection in imagination, b might 
just as well come before a, as a before b. Impressions and 
ideas, according to Hume, differ merely as stronger from 
weaker feelings, nor can any other difference be found 


between them save such as presupposes determination by 
reference to an objective world, which is just what has to be 
accounted for. Of two feelings, which the scientific man 
regards as representing events related to each other in the 
way of cause and effect, so that one can only precede, the 
other only follow, the one which = the effect constantly 
recurs before that which = the cause. The idea of the cause 
is as often suggested by the idea of the effect as vice versa, 
and the idea which suggests the other must come before it. 
Remove the notion of determination by an objective order 
represented by the words ' of the cause ' and ' of the effect/ 
and clearly it becomes indifferent whether cause precedes 
effect or effect cause. 

Thus (a) our ' subjective apprehension of the manifold is 
always successive,' and (b) there is no uniformity in the 
succession ; e.g. ' the apprehension of the manifold in the 
phenomenon of a house which stands before me is succes- 
sive,' just as the apprehension of the positions successively 
occupied by a boat floating down the stream. Yet I judge 
the manifold parts of the house to coexist, and the positions 
of the boat to be necessarily successive. Why this difference ? 
6 Because I may apprehend the parts of the house in any 
order, the positions of the boat only in one, beginning with 
that highest up the stream.' But this will not explain (a) 
why I judge the manifold of the house to exist only in one 
order, viz. together, or (b) why I take the order of apprehen- 
sion to represent the fact in one case and not in the other. 
On the same principle on which I take subjective uniformity 
to represent the fact in regard to the boat, I ought to take 
subjective absence of uniformity to represent the fact in regard 
to the house. The truth is, however, that even in the case 
of the boat, it may very well be the sight (impression) of the 
lower position that recalls the idea of the higher, and if there 
were nothing else than succession of feelings to constitute 
the relation of cause and effect, this would put the relation 
between the positions of the boat on the same footing as that 
between the parts of the house. ' But,' it will be said, ' the 
impression of the lower position, a, can never precede the 
impression of the higher b. No one ever saw a boat at a 
before he saw it at &.' But if the impression only differs 
from the idea as the more from the less lively feeling, what 
difference is there between the sequence of idea on impres- 
sion and that of impression on impression, that the latter 


alone should affect the habit of expectation (according to 
Hume's doctrine), or be regarded as real and objective, so 
that we correct the ' subjective order of apprehension' by it? 

48. A phenomenon, then, if it is to be related to another 
as cause to effect, as uniform antecedent, must be distin- 
guished from ' reproductions of apprehension ' as the object of 
them, and it can only be so distinguished, c if it is subject 
to a rule, which distinguishes it from every other apprehen- 
sion, and which renders necessary a mode of connection of 
the manifold.' ' That which thus renders the connection 
necessary is, according to Kant, ' the unity of understanding ' 
or of ( apperception,' the presence of the thinking subject to 
the manifold, in virtue of which these form one world, and, 
so far as successive, form one succession, so that if a has 
once followed b, it cannot also come before it. Thus even 
the ' subjective order of apprehension ' is, of course, not 
really a matter of chance. The order in which at any time 
I happen to apprehend the parts of a house, like everything 
else that happens, is determined by preceding events ; but 
the order of events being necessarily uniform, the parts of 
the house, to which the order in which they are at any time 
apprehended makes no difference, cannot be an order of 
events. The occurrence of the idea of the boat at position a 
the lower, before the idea of it at b the higher, is as definitely 
determined by preceding events as the actual position of the 
boat at a and the sight of it is determined by its previous 
position at b ; but because the order of events is one, the 
occurrence of the idea at a, b, and c is differently determined 
from the occurrence of the sight. There is no ground for 
this distinction, however, except in the judgment that an 
event, a, which follows another, b, can only follow it. 

This judgment arises from the action of the understand- 
ing in ' applying the order of time to phenomena and their 
existence.' 2 ' If phenomena were things in themselves, no 
man would be able to conjecture from the succession of our 
representations how this manifold is connected in the object. 9 3 
In fact, phenomena=the complex of our representations 
as subject to a rule arising from the unity of apperception. 
Hence the succession of our representations needs only to 
be qualified as one and necessary (a qualification which 
it receives from the 'unity of understanding'), in order to 
become an ' objective connection.' 

1 P. 176; p. 141, Tr. « P. 181; p. 149, Tr. 9 P. 175; p. 143. Tr. 



[Krit. d. r. V. pp. 148-153, 192-205; Tr. pp. 115-119, 161-174.] 

49. The pure conceptions of the understanding may be 
applied either to the ' intuition alone ' (to pure intuition or 
the form of intuition ?) or to the ' existence of a phenomenon,' 
i.e. to the relations of phenomena to each other. In the 
former case they yield ' mathematical principles ' (i.e. 
principles of the possibility of mathematics) ; in the latter 
case ' dynamical principles,' principles of the possibility of 
physical science. The ' objective validity ' of all intellectual 
synthesis depends on the possibility of experience (empirical 
synthesis) corresponding to it; ('the possibility of experi- 
ence is that which gives objective reality to all our a priori 
cognitions ') ; and experience again depends ? on a synthesis 
according to conceptions of the object of phenomena ' (i.e. 
that which our feelings are taken to represent) ' in general,' 
without which we should have no connected whole of ex- 
perience, but only a ' rhapsody of perceptions.' ' This syn- 
thesis, these conditions of the possibility of experience, Kant 
calls 'experience as a priori cognition,' which, he says, 
* possesses truth, i.e. accordance with its object,' only in so 
far as it contains nothing more than the conditions under 
which alone the manifold of intuition becomes a connected 
whole of experience. 2 Hence the ' supreme principle of all 
synthetical judgments,' which Kant states thus (ib.), 'Every 
object is subject to the necessary conditions of the syn- 
thetical unity of the manifold of intuition in a possible 
experience.' The ' principles ' afterwards stated embody 
these 'necessary conditions.' In this part of his doctrine 
Kant has always before him (a) the opposition of analytical 

1 Pp. 151-2; p. 118, Tr. 2 P. 152; p. 119, Tr. 


and synthetical judgments, (b) the definition of truth as 
6 agreement of thought (cognition) with its object.' 

50. l In ' analytical judgment ' we merely ' predicate of a 
conception ' (i.e. of an object thought under certain attributes) 
' that which is already thought in it ' (i.e. certain of these 
attributes). In ( synthetic judgment ' we ' go beyond the 
given conception in order to think in relation with it some- 
thing quite different from what was thought in it.' The 
' supreme principle ' of analytical judgments is that of 
contradiction ; i.e. no such judgment can be true in which 
the object is thought of under contradictory attributes, if 
one of the contradictory attributes ' agrees with the object.' 
This principle is a means for * cognition of truth,' i.e. you 
can ascertain whether an analytical judgment is true by 
asking yourself (if affirmative) whether the contradictory of 
the predicate can be denied of the object. If the contradic- 
tory can be affirmed, the conception does not agree with the 
object, the judgment is untrue. (' Man is mortal ; ' can ' not- 
mortal ' be affirmed of man ? If so, the conception ' mortal ' 
cannot agree with the object 'man,' since contradictory 
attributes cannot belong to the same objects.) But since the 
• object ' in such a case is merely a c thing ' of my own mind, 
certain attributes definitely conceived in unity, it is abso- 
lutely impossible that I should judge an attribute to belong 
to it of which I could also affirm the contradictory. So far 
from being ' of use for cognition of truth,' the principle of 
contradiction is only of use for preventing an error which 
could not possibly occur. There is no meaning, then, in 
calling it a test of the truth of analytical judgment in the 
formal logicians' sense of analytical judgment. It is other- 
wise if ' analytical judgment ' means a process of clearing 
up a confused conception. If the conception of 'man' is 
confused (if you do not quite know what you mean by it, and 
what not), you may be ready to admit propositions about 
' man ' that implicitly contradict each other. Then, as soon 
as the contradiction between them is clearly exhibited, 
according to the law of contradiction you have to reject one, 
and your conception is cleared up. But in such a case, the 
judgments in which you admit these implicitly contradictory 

1 [What is here said on analytical in the lectures on Mill's Logic, §§ 66- 
and synthetical judgments is to be 72.] 
supplemented by what is said below 


propositions are not ' analytical ' in Kant's sense. You do 
not in them ' predicate of a conception what is already thought 
in it,' for you have no clearly articulated conception at all. 
If you had, you could not admit the contradictory propositions. 
Nor is the process of clearing them up an analytical judgment 
in Kant's sense, for you cannot clear them up without 
'going beyond them.' The principle of contradiction, then, 
is not the 'test of truth ' of analytical judgment. It repre- 
sents the law under which you clear away verbal confusion, 
so as to know exactly what you mean by your general terms, 
so as to arrive at those definite conceptions the content of 
which you can state in an analytical judgment. When 
you have arrived at that state in which you can make an 
analytical judgment in the sense of formal logicians (in which 
you predicate of a conception, of which the connotation is 
definitely known, one of its attributes), there is no longer room 
for such a ' test ' as that of the principle of contradiction. 

Thus if analytical judgment means the mental act in 
which you rehearse the contents of a definite conception, 
there is no meaning in calling the principle of contra- 
diction the test of its truth, since it cannot be false. In 
that sense of ' analytical judgment ' in which the principle 
of contradiction can be called its test, it cannot be opposed 
to a synthetical judgment in the sense in which Kant opposes 
it, as that in which we do not ' go beyond ' a given concep- 
tion. The propositions which give a colour to the notion of 
there being ' analytical judgments ' in the sense of formal 
logic are (1) those in which a teacher conveys the meaning 
of terms (those e.g. of a dictionary). But these do not 
represent any process of thought on the part of a teacher, 
and do relate to a matter of fact. No doubt the teacher or 
dictionary-writer must have thought in order to be able to 
give an exposition of the meaning of the name, but this 
thinking is a highly synthetical process, which results in 
the conclusion that, as a matter of fact (e.g.) what is now 
understood by 'gold' is a metal distinguished as follows 
(according to the received scientific account). (2) Those 
which represent the act in which we think of a conceived 
subject under one of its conceived attributes in the process of 
connecting it with something else. There may be real thinking 
represented by ' gold is yellow,' if it represents a stage in 
connecting gold through its yellowness with other objects, 


or of considering whether some newly observed thing is gold 
or not. But this is a synthetic process. 

51. (a) 'All bodies are extended,' is an analytical judg- 
ment according to Kant ; (6) * All bodies are heavy/ a 
synthetical judgment. This cannot mean that ' extension ' 
is included in the meaning of the term 'body/ while 'heavi- 
ness ' is not. Such inclusion is relative to the individual's 
state of mind. To educated men both predicates, to unedu- 
cated neither, are included in what they understand by ' body.' 

Probably Kant means that (a), representing a mere con- 
ception, involves no reference to ' experience,' while (b) does. 
But what is meant by ' experience ' ? Is it meant that the 
predicate ' heavy ' represents sensations repeatedly felt ? 
Then the proposition reduces itself to remembrance, ' I have 
felt a body to be heavy again and again, &c. and never other- 
wise, so that I expect to continue to feel it heavy.' But then 
what does ' body ' mean in such a proposition ? If it represents 
a mere conception in (a), it must do so likewise in (b), and the 
proposition must state the coexistence of such a conception 
with a succession of events in the way of feeling, which is 
nonsense. We may try to reduce ' body ' to a succession of 
feelings in (b), but if so, we must equally so reduce it in (a), 
which knocks up Kant's doctrine as to (a). 

Kant, however, by no means took 'heavy ' to represent a 
feeling or succession of feelings. An intellectual synthesis 
is necessary to give it. ' Weight ' is not a mere feeling, but 
an ' empirical conception,' resulting from the interpretation 
of feeling under the direction of 'synthetic principles of 
understanding ' (in particular the principle of the ' anticipa- 
tions of perception '), and as predicated of body implies the 
conception of the connection of body with the 'whole of 
possible experience.' Undoubtedly, if we had no sensations, 
we never should judge 'bodies are heavy,' but should we 
judge ' bodies are extended ' ? Kant might say ' yes,' since 
extension is a property of pure intuition. But (a) ' sensibility 
alone furnishes us with intuitions,' and though this does not 
mean that sen sation= intuition, it does imply that without 
sensation, as a matter of fact, we should not have intuition ; 
and (6) ' body ' is not pure intuition, so that, though ' pure 
intuition ' does give extension, it does not give it as a predi- 
cate of body. If ' body ' means body as experienced in (b), can 
it mean anything else in (a) ? 


52. The truth is that both judgments are synthetical, in 
the sense that in them thought goes beyond the subject-con- 
ception, which, indeed, would not be a conception if thought 
did not go beyond it. Conception = the thinking of an object 
under relations, and under relations which cannot be isolated 
and summed up, but of which each involves a farther relation. 
Both are analytical as implying analysis of that mere con- 
sciousness of ' something there ' with which our knowledge 
begins, an analysis through which to us the ' something ■ 
becomes the articulated whole which it is in itself. Both, 
again, relate equally to experience. The difference is that 
the correlative analysis and synthesis represented by (a) is 
much more elementary than that represented by (b), so ele- 
mentary that without it there is no definite conception of 
an outward thing at all. It means, ' every body is made up 
of parts outside each other.' In fact, it merely predicates 
of body that which, as predicated of all phenomena, accord- 
ing to Kant himself, is a ' synthetic principle of experience.' 
It is quite true that without extension you cannot think of 
body, but it is misleading to say that in predicating extension 
of body, you do not go beyond the conception of 'body,' 
because the predicate expresses that very act of going beyond 
body, though only to another body, without which body 
cannot be thought of. 

Kant's opposition between the two sorts of judgment is, 
in fact, a survival from the doctrine which opposes what mind 
does for itself to ' facts of nature,' a doctrine upset by the 
admission that s understanding prescribes laws a priori to 
phenomena, and therefore to nature as a complex of all 
phenomena.' It is the same sort of survival, which makes 
him often write as if mathematical truths were only ' sub- 
jectively true,' though their 'objective validity ' is afterwards 
established by the consideration that, nature being consti- 
tuted by understanding, the synthesis of intuitions must 
involve relations constituted by the synthesis of the mere 
forms of intuition. 

53. The notion that the conceived object can be isolated 
— thought of apart from its relation to the whole of experi- 
ence — appears in Kant's explanation of the 'postulates of 
empirical thought.' 1 

According to Kant, if, having conceived an object, I go 

1 P. 192, ff. ; p. 161, Tr. Cf. the statement on p. 193 with that on p. 204. 


on to inquire whether it is merely possible, real, or necessary, 
the asking and answering of these questions in no way 
affects the object, ' object' here meaning a thing thought of 
under a definite complex of attributes, which is neither 
increased nor diminished by the affirmation or denial of its 
reality. 1 

The possibility of mere conception, according to Kant, 
depends on its not involving contradictory attributes. The 
' possibility of such an object as is thought in the concep- 
tion ' is another matter. The possibility of such an object 
constitutes the * objective reality ' of the conception itself. 
It depends either on (a) laws of construction in space, or (b) 
on its being capable of connection with the whole of experi- 
ence according to the ' analogies of experience.' (a) are 
objectively valid ' because they contain a priori the form of 
experience in general ' ; in other words, because 6 the forma- 
tive synthesis hy which we construct a triangle in imagination, 
is the very same as that we employ in the apprehension of a 
phenomenon for the purpose of making an empirical concep- 
tion of it.' 2 As to (b), the question is whether, supposing 
certain uniformities and sequences of phenomena ascertained 
according to these analogies, the conceived object is consistent 
with them according to the same analogies. How can this 
question be answered without further ' determination of the 
object ' ? 

' The principles of modality are not objectively syn- 
thetical.' ' They predicate of a conception nothing more 
than the procedure of the faculty of cognition which gene- 
rated it.' This is true and important, if it means that the 
distinction of possible, real, and necessary is a distinction 
ex parte nostra, a distinction arising from the character of 
our intelligence as in development ; i.e. from the fact that the 
principle which forms the unity of the world (which is also 
the principle in virtue of which I am I) is communicated to 
us, while yet the details, which that principle makes one, are 
not only not communicated to us fully, but never can be, 
since in respect of our animal nature we are among these 
details. It is not that there are three sorts of object, the 
possible, the real, the necessary, but that the real world is 

1 P. 409; p. 368, Tr. Cf. Hume, existent, my idea of Him neither in- 

Treatise, B. i, Pt. in, sec. 7. ' When I creases nor diminishes.' 

think of God, when I think of Him as 2 P. 196 ; p. 164, Tr. 
existeut, and when I believe Him to be 


known to us through a succession of experiences, which 
the unity of the understanding renders a whole of mutually 
qualifying elements, and that thus to us in any stage of 
experience there are many possibilities of which we cannot 
say yet whether they are real ; they are possibilities, as not 
being inconsistent, according to the formal conditions of 
experience, with our hitherto experience, but possibilities of 
which we cannot say that they are real because our hitherto 
experience is only a part of possible experience. 

54. Kant's error (I think) lies in treating such possibility 
as ' objective possibility.' It is not objective possibility, un- 
less consistent with the whole order of the world as it is, and 
whatever is possible in this sense is also real. In this latter 
(the true) sense of the ( objectively possible,' it is quite true 
that the object, when from being possible it becomes real, is 
* not farther determined,' but only so because in this sense 
the possible and the real are the same. To the objectively 
possible in the above sense, the occurrence of a sensation (a 
new perception) on our part makes no difference. To the 
subjectively possible it may make a great difference. It may 
verify or falsify an hypothesis. A ' subjectively possible ' 
conception must precede every experiment. The experiment 
shows whether a relation of phenomena, supposed to be 
possible, is real or not. Through it nothing becomes real 
that was not real before. c Is it not the case, however,' it 
may be said, ' that through it what was conceived as possible 
comes to be conceived as real, and that without any change in 
the content of conception?' No, because the experiment 
always involves the analysis of some phenomena not analysed 
before ; it enables you to judge that a really always accom- 
panies b, whereas before you only guessed it, because after a 
crucial experiment you are able to set aside all conditions in 
the complex phenomena, which included b and which a had 
been found to follow, except b itself. 

Thus taking the ' possible object ' in one sense, it is quite 
true that the occurrence of a perception corresponding to it 
makes no difference to its content; but of such an object it is 
unmeaning to say that, through the occurrence of perceptions, 
from being possible it becomes real. Taking ' possible object ' 
in another sense, it is quite true that the occurrence of a 
perception converts its possibility into reality, but in doing 
so, it further determines the conception of the object. 



[See especially Krit. d. r. V. pp. 197-200; pp. 166-169, Tr.] 

55. The distinction between outer and inner sense can- 
not be explained (1) as a distinction between consciousness 
due to ' external ' and that due to ' internal ' stimulus, be- 
cause (a) nervous stimulus cannot be distinguished into 
outer and inner ; and (b) because ' outer ' sense means the 
sense of an outer object, e inner ' the sense of an inner 
object, and the nervous stimulus of consciousness is not the 
object of consciousness ; nor can it be explained (2) as the 
distinction between ' ideas of sensation ' aud ' ideas of reflec- 
tion,' between ' impressions ' and ' ideas,' between perception 
and memory or imagination, because (a) to any conscious- 
ness of an ' outer ' object the second member in each of these 
pairs is as necessary as the first, and (b) the object of con- 
sciousness, as such, may be just as much outer when there is no 
sensation present as when there is. When I remember my 
house, the object of consciousness is as much c outer ' as when 
I am looking at it. A remembered pleasure is not an out- 
ward object, but no more is a pleasure at the time of being 
experienced. Thus the difference between outer and inner 
sense lies in the relation of the object as an object of con- 
sciousness, not in consciousness as apart from the object. 
The relations of the house on the one side, of the pleasure 
on the other, as objects of consciousness, being the same 
whether the house is perceived or imagined, whether the 
pleasure is being enjoyed or remembered, the difference 
between perception and imagination, between enjoyment 
and recollection, is not a difference between inwardness 
and outwardness of the ' sense.' We are thus brought to 
adopt as the only tenable distinction between outer and 
inner sense that between the consciousness of objects as 
related to each other, not to the conscious subject, and 
the consciousness of objects as changes in the state of the 

VOL. II, p 


conscious subject. Thus it is really a distinction, not between 
two sorts of sense as such, but between two sorts of intel- 
lectual interpretation of sense, two functions of the under- 
standing in the connection of phenomena. 

56. In most passages, at any rate, Kant uses ' internal 
sense' for the consciousness of changes as in oneself, in the way 
described. It is this which, as ' empirical apperception,' or the 
' consciousness of self according to the determination of our 
states in internal perception,' he opposes to ' transcendental 
apperception.' l It is of this, again, that he is thinking when 
he shows 2 that % internal experience is possible only mediately 
and through external experience,' because ' consciousness of 
my own existence as determined in time ' implies a ' perma- 
nent something external to me,' in relation (contrast) to 
which alone a consciousness of time is possible, and which 
cannot be one of my ( representations,' because, if it were, it 
would be in time, and thus not permanent. This is Kant's 
reply to 6 problematic idealism,' which he affiliates to Des- 
cartes, the doctrine that the only immediate certainty lies in 
the consciousness of one's own existence, and that the exist- 
ence of the outer world is known mediately or inferentially 
through this. Kant retorts that in order to such conscious- 
ness, as of the succession of my inward states, there must 
already be knowledge of that which is supposed to be inferred 
from it, viz. of a permanent something other than these 
states, in order to render consciousness of their succession 

This outward something, however, as Kant understands 
it, cannot be c outward ' according to the ordinary mean- 
ing of the term. For, according to the rest of his doctrine, 
it must result from the determination of phenomena (the 
modifications of sensibility) by the 'unity of apperception' 
or the principle of this. The ' permanent something,' accord- 
ing to him, it is true, which must be known in order to 
render the consciousness of succession possible, cannot be 
the ' transcendental ego ' itself, because this is not knowable 
from the want of a corresponding phenomenon. It must be 
what (in the account of the first ' analogy of experience ') he 
calls ' substantia phenomenon.' It is very difficult to make 
out how he understands this. It is the 'permanent' or 
s real ' in phenomena, but implies determination of these by 

1 P. 572; Mahaffy's trans, loo. cit. p. 200. > P. 199 ; p. 168, TV. 


the category of substance, a particular ' function ' of the 
unity of apperception. 

57. Now Kant is quite right in saying that 6 inner sense,' 
as consciousness of successive modifications of one's state, 
implies outer experience, the conception of permanent objects 
other than such modifications. It is a mistake of ' meta- 
physicians ' to say that we are primarily conscious of our- 
selves, in the ordinary sense of the words, or of our states 
as ours. In order to such consciousness we must already 
have been conscious of objects, neither as distinctly outer 
nor distinctly inner, and have gradually come to distinguish 
our own changes from what we suppose to be permanent in 
them. It is true that only through modifications of sensi- 
bility, determined by the presence to them of the thinking 
subject, are we conscious of objects at all. But in the order 
of our experience, the consciousness of objects precedes re- 
flection on the conditions of their presentation as ex parte 
nostra ; precedes and determines the ' inner sense,' according 
to the meaning which we have so far attached to it. We 
see things before we are conscious of the sensation of sight 
as a sensation. To have a sensation is different from being 
conscious of having a sensation; and not only so, but that 
determination of sensations by the self-conscious subject (the 
subject present to and distinguishing itself from all) which 
is necessary to their becoming the sense of objects, is different 
from the consciousness of having sensations. 

58. The question, however, arises, (1) whether the sense 
of objects, as preceding such e inner sense,' is properly con- 
sidered ' outer ' ; whether the distinction of ' outer ' and 
' inner ' does not represent a process of reflection subsequent 
to the consciousness of objects, and of which the two sides 
are strictly correlative, neither being prior or sequent to the 
other ; (2) whether, when time is called the form of ' inner 
sense,' there is not a confusion between modifications of 
sensibility (properly neither outer nor inner, and of which 
alone there is any case for calling time the form), and con- 
sciousness of such as of changes in my state in opposition to 
objects other than me, which is properly c inner sense,' but of 
which time is no more specially the form than of any other 
consciousness of change. 

Looking to the consciousness of a c permanent some- 
thing ' which is the condition of the consciousness of my 

F 2 


states as successive, there is nothing in either of the factors 
from which it results, either in the phenomena, modifications 
of sensibility, which in themselves are neither outer nor 
inner, or in the c unity of apperception,' to qualify it as out- 
ward. It only becomes ' outward ' in so far as consciousness 
of changes as in my state is awakened and opposed to it. 
As an antecedent condition of such consciousness, it is not 
outward any more than inward. 

As to question (2), it is clear that time is the form of all 
change. There is no propriety in calling it the form of 
inner sense according to the above meaning, unless it be true 
that changes are only in time so far as reflected on as suc- 
cessive modifications of my state, which was not what Kant 
meant. He regarded ' time ' as the form of all phenomena, 
as modifications of sensibility, and only came to speak of it 
as a form of inner sense from the confusion of such modifica- 
tions with the consciousness of them as changes in me in 
opposition to changes in things. 

59. His double usage of ' internal sense ' appears on 
pages 127-129 (Tr. pp. 93-95). He first uses * internal sense ' 
for that which 'represents to us our own consciousness, 
only as we appear to ourselves, not as we are in ourselves 
. . . ourselves only as inwardly affected ; ' in short, * em- 
pirical apperception.' But he proceeds, ' That which deter- 
mines the internal sense is the understanding,' &c. But it 
is only the determination by the understanding which yields 
1 internal sense ' according to the above meaning. The text 
implies that there is first internal sense, and that then it is 
determined by the understanding. But without such de- 
termination there is no sense of objects at all, much less a 
sense of objects as distinguished into outer and inner, since 
without it there is no ' synthesis of the manifold.' 

In truth, the ' internal sense ' of the paragraph beginning 
' That which determines,' &c, and of the following, is not 
the ' empirical apperception ' of the preceding paragraph, 
but simply the sensuous as a manifold in time, which 
according to Kant all the sensuous is, though somehow some 
of it is also a manifold in space. Kant then uses ( inner 
sense ' in two meanings : (1) for the sensible as a manifold 
in time, which all the sensible (according to him) is; (2) for 
' empirical apperception.' In the second meaning, according 
to his own snowing, it presupposes the action of the under- 


standing. If confined to the first meaning, (a) there is 
impropriety in calling time a form of inner sense, taking 
' sense ' apart from the determination which Kant ascribes 
to understanding ; there is nothing to distinguish it as inner 
except the form of time itself ; ' inner sense ' = sense in time. 
'Time is the form of inner sense,' then=' time is the form 
of sense as in time.' (b) It does not appear how, space and 
time being alike forms of the sensible apart from the action 
of the understanding upon the sensible, while time is the 
form of all sense as such, space should yet be the distinguishing 
form of a certain sort of sense. Kant says, ' The internal 
sense contains merely the form of intuition, but without any 
synthetical conjunction of the manifold therein ; - but if it 
'does not contain any determined intuition,' how can it 
contain the ' form of intuition ' ? To ' contain the form of 
intuition ' must mean that it contains the relation of succes- 
sion, the relation in virtue of which the data of sense are 
one before, one after, the other. No doubt it really contains 
this relation, but only through the synthesis effected by 
1 apperception.' Kant immediately afterwards says that the 
conception of succession is derived from motion, as i an act 
of the subject ' by which it ' determines the internal sense 
according to its form ; ' and ' such synthesis of the manifold ' 
(as that which yields the conception of succession?) 'the 
understanding does not find in the internal sense, but pro- 
duces, in that it affects this sense.' It is admitted, then, that 
only an act of understanding can constitute that ' synthesis 
of the manifold ' which is necessary to the conception of 
succession. The question is whether there could be succes- 
sion, time, or ' form of inner sense ' itself, without such 
synthesis. Kant seems to have thought there could be, and 
hence speaks of time as belonging to the determinable, or 
as itself the determinable, prior to the act of determination 
by the understanding. 1 In fact, he never gave up the 
notion that the sensuous or determinable (what has yet to 
be determined) is given to the understanding under two 
distinct modes of multiplicity, as a manifold in space and 
a manifold in time. Though considering it merely deter- 
minate, he yet assigns to it such actual determination as 
relation in space and relation in time constitute. 

1 P. 130, note; p. 96, Tr. 


60. 1 There is the distinction (a) between sensation and 
perception ; (b) between perception and memory or imagina- 
tion ; (c) between sensible objects as successive, and sensible 
objects as coexistent and mutually limiting. 

When Kant calls time the form of inner sense, space of 
outer, he has in view the last distinction, but he speaks as 
if there were some other distinction than this between 
objects of outer and inner sense — a distinction conveyed by 
these terms ' outer ' and ' inner ' — and as if the distinction 
between the determination by the form of space and the de- 
termination by the form of time were found to go along with 
this independent distinction between objects of outer and of 
inner sense. He does not treat the distinction between the 
form of space and the form of time as constituting that 
between outer and inner sense, but the latter as existing on 
independent grounds, and the former as happening to coin- 
cide with it. 

On examination we find (1) the distinction between outer 
and inner sense untenable in itself, and (2) that, if we take 
it — under a protest against the use of the terms ' inner ' and 
* outer ' — to represent (a) or (b) above, then distinction (c) 
will not coincide with it. 

There is (d) the distinction between the consciousness 
of objects as given — whether perceived or remembered, 
whether given through the senses of sight and touch, or 
those of hearing and smelling — and the consciousness of 
changes of my state implied in the giving of such objects. 

In itself, the presentation of an object is neither outer 
nor inner (or, if we like, both outer and inner). It is the 
presentation of an object through being a change of con- 
sciousness. It has its outer side in the relation of the 
object to other objects, its inner side in the relation of the 
modification of consciousness to other modifications of con- 
sciousness (the latter being the condition of the former). 

When objects have been presented, however, attention 
may be directed to the modification of consciousness implied 
in the presentation, and attention so directed may be called 
' inner sense.' 

It may be held that such ' inner sense ' yields first the 
consciousness of mere time, as a relation (though not of 

1 [This section, which is partly a resume of the preceding five, is from a 
detached sheet of manuscript.] 


objects related in the way of succession). This, however, is 
not what Kant meant in calling it the ' form of inner sense.' 
He meant that it was a relation between changes of con- 
sciousness ; but these are not really different from the pre- 
sentation of objects through them, which he called ' outer 
sense. 5 

Time exists for consciousness in presenting its own 
modifications to itself. 

Space exists for it as presenting objects under a condition 
which is the reverse of that under which its own modifications 
are presented. 1 

1 [Compare what is said on ' outer and inner sense ' below in the lectures on 
Mill's Logic, sections 98-101.] 


[See especially Krit. d. r. V., pp. 346-350; pp. 307-310, Tr.] 

61. The objection to the Kantian doctrine of space and 
time, which presents itself to every one, is that the ' empirical 
reality ' which he allows them is not enough. As he expressly 
says, it is a reality conditioned by forms of human sensibility. 1 
How then conld there be space and time when as yet there 
was no such sensibility? Is not that according to Kant 
equivalent to asking how there could be time before the 
possibility of time ? Yet does not all science show that there 
were (to say the least) bodies moving, i.e. successively occupy- 
ing different places, and of which the conditions changed, 
long before there was man? whereas Kant tells us that 
changes presuppose time, 2 and motion presupposes both space 
and time. 

62. Time, — Kant is quite right in opposing the notion 
that things qualified by relation are given independently of 
thought or the mind, and that we then proceed to abstract 
relations, and among them, as the most universal, that of 
time. But on his own part he is wrong in putting the case 
as if time were first given by the mind as a c form ' or relation 
(' as that which effects that the content of phsenomena can 
be arranged in a certain manner'), and that then objects are 
fitted into this form. ' Time ' is an abstraction (a relation 
abstracted from the related objects in which reality lies), 
though not an abstraction from a world given independently 
of thought. 

It is a mistake to speak as if time rendered change 
possible : the same sort of mistake as to say that separation 
of interests renders the state possible. When you come to 
analyse what is involved in the existence of a state, you find 
that if all interests were identical, there would not be a 
state. On the other hand, the state tends to overcome, and, 

1 Pp. 61-62 ; p. 26, Tr. 2 P. 69 ; p. 32, Tr. 


so far as it approaches perfection, actually overcomes, sepa- 
ration of interests. In other words, just so far as there is a 
state, interests no longer are merely separate. In the state, 
separation of interests may be said to exist as one factor of 
-fe) reality, but as in one sense neutralised by the other 
factor, which is its opposite, viz. the sense of common in- 
terest. Neither would be what it is without the other, but 
in the state neither retains any separate reality. So in real 
change, time, as a mere relation in virtue of which this is 
over before that begins, has no real existence. Let a process 
of change be represented as states b, c, d, &c. of something 
other than the states, which shall be called A. State b 
determines c, and c, as determined by b, determines d ; so 
that b and c have not really ceased to exist in the existence 
of d ; and A — that of which they are all states — continually 
determines or exists in all. Here, then, is no mere or abso- 
lute before and after. The relation of time is involved in the 
reality of change, but only as one factor of the reality, of 
which the other is its opposite, viz. the qualification of the 
state existing at one moment by states existing at other 
moments in virtue of a law or subject equally operative in 
or constitutive of all. Except as neutralised by this opposite 
factor, time has no reality ; it is a mere abstraction. In 
short, there is really no such thing as mere time ; it only 
exists for our abstracting intelligence. Let us come then to 
real change, or motion, or the succession of events in a defi- 
nite universe where each is qualified by all, the world of 
becoming. (1) Is such a world possible except for a think- 
ing subject ? (2) If not, is there for such a subject what 
we mean by time ? (3) For such a subject, or in its reality, 
has the world of becoming a beginning and end ? 

63. It has been sufficiently shown that the c cosmos of 
our experience ' is only possible in relation to a thinking 
subject, as that for which appearances, past as feelings, are 
present as facts determining and determined by all others. 
' This,' it may be said, ' may be true of the cosmos of our 
experience, but how can it be true of that which is not phe- 
nomenal, as must have been the world (the series of events) 
that preceded sentient life ? ' The answer is that it is not 
our sentience that is the condition of there being for us a 
phenomenal world, though the fact that we are sentient (and, 
so far, merely parts of this world) limits (renders inadequate) 


the mode in which we understand it, i.e. in which it exists 
as a phenomenal world for us. The condition of there being 
for us such a world is the existence of a reason, which we 
call ours, but which we cannot suppose, without hopeless 
contradiction and confusion, to have begun with our sen- 
tient life, any more than we can suppose the principle in virtue 
of which we say c we ' or ' ours 9 to have begun with that 
life. There is no meaning in speaking of a series of events, 
* revealed to us by science ' as antecedent conditions of life 
and sentience, which must have taken place when as yet life 
and sentience were not, as if they did not belong to our phaj- 
nomenal world, 'the cosmos of our experience.' They are 
onditions of what we experience, determined just as much 
by relation to what we experience as it by relation to them. 
Limit e our experience ' to the succession of our feelings, and 
there is no ' world of experience.' Extend it so as to mean 
that which determines our feeling, and it must include con- 
ditions antecedent to the appearance of sentient life just as 
much as any other. If e science ' reveals such conditions, 
the right inference to draw is, not that the world is inde- 
pendent of thought, but that thought, the condition of there 
being such conditions, does not come into being as a de- 
velopment of life and sentience. 

64. Admitting, then, an eternal thinking subject, as the 
correlatum of nature, without which nature could not be, 
what is nature for such a subject? The answer is, it is just 
what it is for our reason, which is this eternal thinking 
subject. 1 It is not essential to there being a nature for us 
that we should be sentient. 2 Facts of sense exist for us as 
understood, or as constituents of a nature, when no longer 
felt. Nor, as merely felt, are they facts for consciousness at 
all. The circumstance that we not only know what facts of 
feeling are, but ourselves feel, so far interferes with our 
knowledge. For reason (and, except for reason, there is no 

1 In us the function of reason, as that we are parts of nature; but just so 
rendering a nature possible for our con- far as we are parts of nature there is 
sciousness, is not its sole function. It no nature for us, i.e. as the object of 
renders morality possible too. And if our consciousness. It is not as sentient 
we speak of the eternal subject as (rod, that we are able to present nature to 
we must not suppose, because God ren- ourselves as an object, nor, if there 
ders nature possible, that this is the were nothing but successive feelings, 
full account of God. He must at least would there be a nature at all. But it 
render morality possible too. is equally true that without feeling 

2 It is through our sentience (and there is no nature, 
the life on which sentience depends) 


nature at all), nature is a system of becoming, which rests 
on unchangeable conditions. Subject to these conditions, 
new events take place. The state of the natural world to- 
day is what it never exactly was before. But it had its 
necessary possibility in the state of yesterday, just as that 
possibility has its necessary realisation in the state of to-day. 
Such is nature for that thinking subject which renders nature, 
as an object, possible for us, and there is no reason to suppose 
another such subject for which it would be anything different. 
No doubt nature in other ways is wholly different to us from 
what it would be to a being that was not, as we in virtue of 
our animal life are, part of it. We understand that it is the 
system described ; but to us, as we are at any moment of our 
lives, all nature but a little part is ' expunged and rased.' 
Nature brings each man at every moment his own joy or 
sorrow, which is no one else's. Save in respect of the formal 
conditions of knowledge, every one finds nature different from 
what every one else finds it; much more, except in that 
respect, must it be different for each man on the one side, 
and on the other for the thinking subject in its full reality, 
which must be determined by relation to the whole of nature. 
Thus, though nature is really, or for the eternal thinking 
subject, for God, what it is for our reason (i.e. for this subject 
as enabling us to present a nature to ourselves), when we 
come to say what it is for our reason, we cannot get beyond 
the mere formal conditions of there being a nature at all. 
We do no more than state these when we give such a formal 
definition of nature as the above. 

65. This definition immediately suggests the question, 
Has nature (the system of becoming) a beginning and end ? 
and if so, what? A beginning of nature would be an un- 
conditioned occurrence 1 ; an end of it would be conditions 
that had no effect. A beginning and an end of nature thus 
alike involve contradictions. In another sense, indeed, nature 
may be said to have beginning and end, each being God, 
since the thinking subject is the condition of its possibility, 
and, as yielding man who shares the divine consciousness, it 
returns to God. But the process of nature, in yielding man, 
does not come to an end as a process of becoming. 

1 The 'primitive matter' of the evo- nothing, or else is so conditioned as to 

lution theory, -which contains the ' pro- be virtually already all that is * evolved ' 

mise and potency of all forms .of life,' is from it; but if so conditioned, it is not an 

either nothing at all, and can explain absolute first, not unqualified matter. 


Are we to hold (a) that the world of becoming has begin- 
ning and end in time and space ; (b) that it has no beginning 
or end ; or (c), if we find it equally impossible that it should, 
and that it should not, have beginning and end, are we to 
hold that this equal impossibility belongs to its real nature, 
or (d) that it is due to the infirmity of our thought ? 

It is the condition of every event that it has an antece- 
dent event. There can then be no stoppage in regress from 
event to event, for any first event would be indeterminate, 
would be nothing. On the other hand, a determination of 
any event (of any now) by an indeterminate series of events 
would be no determination at all. For the same reason, then, 
for which we deny that there can be a first event, we must 
deny that events form an endless series. 

66. Kant would say that, though this contradiction does 
not affect or relate to ' things in themselves,' it is inherent in 
the nature of empirical reality or the world of experience ; 
and when we have given this ' world of experience ' the full 
extension which his theory logically requires for it, there 
remains nothing outside it, nothing to be a ' thing in itself,' 
but the unconditional thinking subject itself, which is 
the source of the categories. 1 The 'world of experience '= 
sensibility as related to reason, and in consequence of that 
relation determined by the categories, ' conceptions of the 
understanding ' which connect all ' modifications of sensi- 
bility ' (phenomena) with each other, so that they become a 
changing world. The contradiction between having a be- 
ginning and not having one is a necessary incident of this 
world. The unconditioned subject in relation to feelings 
renders them changes, for which a beginning must always be 
sought and can never be found, or — to vary the expression — 
of which it is the nature to involve the contradiction of being 
each determined by an indeterminate series. Thus the con- 
tradiction involved in ' our conception ' of change arises out 
of the relation of reason to sensibility as constituting the 
knowledge of nature ; but, according to Kant, this relation 

1 Thus in itself, as distinguished nature which results from the relation 

from its work in constituting the world of reason to sensibility. The Kantian 

of experience through relation to sensi- answer would be that the fact of 

bility, what right (it may be asked) morality — of action at least affected by 

have we to ascribe any such separate the judgment 'I ought' — shows it to 

existence to this subject? Its reality have another function than that which 

must lie in its function— i.e. in the yields nature. 


also constitutes the * empirical reality ' of nature ; therefore 
the contradiction belongs to this reality. 

We then come to the difficulty how the relation of 
reason to sensibility can constitute the reality (in any sense, 
1 empirical ' or other) of what took place before there was 
human sensibility. 

In regard to events which took place before there was 
man, are we, from Kant's standpoint, to suppose (a) that, 
though prior, they come into being so far as man finds them 
necessary to explain the phenomena of his experience, i.e. 
that their very prior existence is made by knowing man ? or 
(b) that their relation to a future human sensibility was the 
condition of their prior existence, as an existence for intelli- 
gence ? or (c) that before man was they existed in relation 
to a present sensibility, a sensibility of which that in each of 
us is in some way a reproduction ? 

67. Kant certainly uses language which seems like view 
(a). 'In the transcendental aesthetic,' he says, ' we proved 
that . . . all objects of a possible experience, as presented to 
us — as extended bodies or as series of changes — have no self- 
subsistent existence apart from human thought.' ! This may 
be admitted, and yet it may be held that ' human thought, 
sensibility qualified by reason, does not begin with the 
* appearance of man upon the earth ; ' that it, and with it the 
possibility of nature, exists eternally in God ; so exists as a 
mode of the spirit, in freedom from the delusions which 
arise in each man, (a) from his knowing no other mode of 
the existence of spirit than that which constitutes nature, (6) 
from the limitation in his knowledge even of this — from the 
limitation of his experience. 

On the other hand, two passages 2 would seem to imply 
that objects which we may come to discover either as now 
existing in space, or as having existed in past time, only so 
exist in virtue of the discovery, only exist when discovered. 
These passages look like a virtual admission of Berkley anism. 

1 P. 346, ff ; p. 307, Tr. 

2 ' That there may be inhabitants in in so far as I can represent to my own 
the moon, although no one has ever mind that a regressive series of possible 
observed them, must certainly be ad- perceptions . . . conducts us to an 
mitted ; but this assertion means only elapsed series of time as the condition 
that we may in the possible progress of the present time ' (p. 349 ; p. 309, 
of experience discover them at some Tr.). (Kant leaves unexplained the 
future time ' (p. 348 ; p. 308, Tr.) difficulty how a ' progress of experi- 
' The things that really existed in past ence ' becomes also a ' regressive series 
time ... are to me real objects only of perceptions.') 


(esse = per dpi), qualified merely by a recognition of the work 
of thought in constituting the connection between sensuous 
representations, without which there are no perceived objects. 
He admits, indeed, a 'transcendental object of experience ' in 
which these empirical objects, which we come to discover, 
may be said to be given, and which itself ' exists prior to all 
experience ' ; but this, he explains, is merely the ' intelligible 
cause ' looked at in a particular way, viz. as corresponding 
to our receptivity ; and one cannot find any consistent view 
of such ' intelligible cause ' in Kant except as the ' transcen- 
dental subject' which conditions the connection between 
phenomena. To say that they are given in the c transcen- 
dental object ' is either a mere fiction of ours, or it must mean 
that they are given in the transcendental subject. In this, 
however, in so far as it is related to a sensibility, either they 
are not really ' given ' at all, except as gradually discovered ; 
or, if we say that we only discover what is already given to be 
discovered, that sensibility, which in relation to reason is the 
condition of their existence as discovered, must also be the 
condition of their existence as ' given.' 

We cannot explain their existence as ' given ' — their 
existence before the progress of our experience yields them to 
us — on the supposition (b), i.e. of a relation, existing for an 
intelligence without sensibility, to a future sensibility. Such 
a relation is really a contradiction in terms, for the sensibility 
supposed, if any relation to it is to exist for the supposed 
' pure ' intelligence, must be present to such intelligence ; 
nor, since sensibility is a condition of there being time, could 
the relation expressed by ' futurity ' exist for an intelligence 
unrelated to sensibility. 

As a series of perceptions ex parte nostra, that which 
' conducts us to an elapsed series of time as the condition 
of the present time ' is a progress, not a regress. Just as, if 
we take Hume's doctrine of cause and eifect strictly, — i.e. 
take it as reducing the relation to mere sequence of our 
feelings, — we are met with the difficulty that the feeling 
called effect occurs to us before that called cause as often 
as otherwise (e.g. 'fire causes warmth '= the sight of fire 
always precedes the feeling of warmth, but it is a mere 
chance in any case whether I see the fire before I feel it, or 
vice versa), so it would be with Kant's doctrine if we took 
it to mean that the reality of past events consists in our 


coming to discover them. We must either understand the 
* progress of experience ' to mean a succession of phenomena 
issuing in some given perception, as regarded from a sup- 
posed point of view in the past from which it appears as a 
progress, — i.e. not as a process ex parte nostra, — or, if we 
take it as a process ex parte nostra, we must say that it is 
one in which we retrace the real progress from phenomena 
to phenomena, and that, as such a retracing, it is a regress ; 
a progress, indeed, on our side from a given perception to a 
later perception, but a regress as from the given perception 
to really earlier phenomena. Such progress or regress will 
not constitute, but only reproduce for us, the reality of the 
past pheenomena ; and this ' reality of past phenomena,' if 
' phenomena ' means anything, implies a sensibility to which 
they were relative, other than, though reproduced in, the 
series of perceptions ex parte nostra. 

68. (c) is the only tenable view. Sensibility is the con- 
dition of existence in time, of there being events related to 
each other as past, present, and future. Ask yourself what 
meaning the terms ( now ' and ' then ' have except as derived 
from a relation between a perpetually vanishing conscious- 
ness and one that is permanent, and you find they have none. 
' Time ' is simply the relation between any ' now ' and e then.' 
When we say that there was time or there were events in 
time before man began to exist, we mean that there were 
events, of which each was thus related to another as ' now ' 
to ' then.' When we speak of time that has elapsed between 
one event and another — between, e.g., the battles of Marathon 
and Waterloo — we mean that a definite number of periodic 
changes, each consisting in the departure of the sun from, 
and its return to, a certain apparent position relatively to the 
earth, have taken place in the interval. When we speak of 
a lapse of time, ' millions of years,' before the formation of 
the solar system, we mean, I suppose, that between some one 
point in it, a, and another, b, events took place, during the 
course of which, if they had happened under the solar sys- 
tem, so many revolutions of the sun (or earth) would have 
been completed. But all these expressions about ' events ' 
and ' happening ' and ' taking place ' imply or derive their 
meaning from a sensibility, of which the perpetually vanish- 
ing modes are held together by a subject equally present 
to, and distinguishing itself from, all of them. 


Thus it appears that changes, and time as the relation of 
before and after between changes, presuppose a sensibility 
determined by reason. Just as each man in fact can only 
think a past before he was by throwing back his sensibility 
(' Tf I had been there, I should have seen it '), so the possi- 
bility of changes prior to the existence of feeling on earth or 
anywhere else must have lain in a sensibility which never 
was not, in such sensibility as is related to a self-distinguish- 
ing subject. Such sensibility is the eternal condition of 
time. Out of it, in relation to reason, arises the eternal fact 
of change or ' nature,' carrying with it the contradiction of 
determination by endless antecedents. (As Aristotle said, 
iraaa <pvai9 vXrjv zyzi ; there must always be a residuum of 
unintelligibility in nature; it would not be nature otherwise.) 
Such f nature ' is at once not God, and that without which 
God would not be what He is. Not God, because reason, 
which in relation to sensibility constitutes nature, exists in 
other modes also. Its action in us, as distinguishing itself 
from nature, and, through such distinction, issuing in art and 
moralit}^, may show us this. 1 What more than the condition 
of the possibility of nature God is, only man's achievement 
in art, morality, and religion enables us to say. 

69. To return, then, to the three questions asked above, 2 
the answer to (1) is, that nature, or the world of becoming, 
is not possible except for, or in relation to, a thinking sub- 
ject ; to (3), that for such a subject (in relation to an eternal 
sensibility) nature involves the contradiction of determination 
by an indeterminate series — absolute beginning can neither 
be affirmed nor denied of it; to (2), that for such a subject, 
in the same relation to an eternal sensibility, ' time ' exists, 
as involved in change, though not in the abstraction which 
is a fiction of our human intelligence. 

My excuse for this ' metaphysical ' treatment of these 
questions is that men of science do not help us about them. 
They sometimes talk as if science proved that the world 
must come to an end and must have had a beginning (which 
again they sometimes tell us must have been ' an arbitrary 
fiat of a creator,' which they leave a mere phrase) ; but on 
examination one finds that they do not mean absolute begin- 

* Art, whatever else it implies, im- this is self-distinction from nature, 

plies a joy in nature which impels to 2 [Section 62.] 

representation of it ; and the condition 


ning* or end, but merely transition into or out of a state in 
which the ascertained laws of matter are applicable to it. 
Nor does science seem able really to account for the appear- 
ance of sensibility on earth or elsewhere, in such a way as to 
dispense with the supposition that this appearance is a mani- 
festation in time of its eternal existence as the condition of 
there being time. 





Note of the Editor. 

Of the following discussions on Kant's Moral Philosophy the main por- 
tion (K, L, and N) formed part of Green's first two courses of professorial 
lectures, delivered in the summer and winter of 1878. The subject of these 
courses was what he understood as the ' Metaphysic of Ethics,' i.e. (1) the 
inquiry into the determination by motives as that which constitutes freedom, 
in the sense in which all moral action is free; (2) the inquiry into the 
relative worth of motives, or the conception of the chief good. Most of 
these lectures were subsequently embodied in his Prolegomena to Ethics, 
but they contain a fuller statement and criticism than is there given of the 
cardinal points in Kant's theory, and these portions are printed here. The 
date of M is fixed by a reference to the 2nd edition of T. K. Abbott's work, 
Kant's Theory of Ethics, as not earlier than 1879. The MSS. from which O 
and P are taken belong apparently to the same group as those which were 
used for the professorial lectures of 1878. 

The references, as before, are to Hartenstein's edition of Kant's works, of 
which vol. iv. contains the Grundlegung zur Metaphysik der Sitten, and 
vol. v. the Kritik der praktischen Vernunft : the translation referred to is 
that of T. K. Abbott in the work mentioned above. 




70. What is the province of moral philosophy? 'It 
deals with man as a moral agent,' but what is moral 
agency? Is there anything about it to distinguish moral 
philosophy from natural science ? There is an anthropology 
which is simply a branch of natural science. It regards 
man, like any other animal, as a mere result of natural in- 
fluences ; inquires how by a long course of adaptation to 
environment the human animal has been so modified as to 
be what he is ; what are the chief varieties of this animal, 
and how they are to be accounted for. There is no doubt 
that anthropology so understood is a valid science. Is 
moral philosophy merely a branch of it ? an inquiry into 
certain secondary modes of pleasure and pain, arising from 
adaptation to a social environment, which determine the 
actions specifically called moral, and into the bearing of 
such actions upon the further natural production of plea- 
sure? If so, it is a purely natural science, moral agency 
being merely a most complicated form of natural agency, 
complicated by the development of the social * medium ' or 
'organism,' and its reaction upon the individual. 

71. On the other hand, we may draw an absolute distinc- 
tion between moral and natural agency. The formula for 
such distinction is best given by Kant : ' Everything in 
nature works according to laws ; the distinction of a rational 
being is the faculty of acting according to the consciousness 
(' Vorstellung ') of laws, i.e. according to principles.' These 
laws may be merely natural laws ; still an agent determined 

a 2 


merely by them (or, more properly, by another agent accord- 
ing to them) is essentially different from an agent determined 
by the consciousness of them ; and if man is an agent of 
the latter sort, there will be place for inquiry, quite distinct 
from natural science, into the forms of 6 inner life ' arising 
out of this consciousness, an ' inner life ' to which the index 
will be the language and institutions of men ('language,' 
not as the mechanism of speech, but as its content, that 
which is said). 

[After an argument substantially the same as that of 
the Prolegomena to Ethics, the conclusion is arrived at, 
' that the experience of man as distinguished by unity 
of consciousness, or self-consciousness, or general concep- 
tions, or conception of laws, forms a distinct object of 
inquiry, with which * anthropology ' as a physical science 
cannot deal.' Of this consciousness there are ' two parallel 
activities (parallel in the sense that they do not meet, can- 
not be brought to a common point) ; one yielding nature 
and the sciences of nature, the other yielding the moral life. 
Corresponding to these two exercises of reason are two forms 
of philosophy, each distinct from any kind of natural science. 
The object of one is to answer the question, What conditions 
on the part of consciousness are implied in the fact that there 
is such a thing as knowledge, or that a ' cosmos ' arises in 
consciousness ? of the other to answer the question, What 
are the conditions on the part of consciousness implied in the 
fact that there is such a thing as morality ? ' The lecture 
then continues as follows.] 

72. The view here stated, of the distinction between the 
natural sciences (or sciences properly so called) on the one 
side, and the inquiry into the functions of reason as (1) 
theoretic and (2) practical, is founded on that of Kant. Nor 
is it possible to discuss the present state of the question in 
regard to the possibility of moral philosophy, as distinct from 
a natural science of anthropology, without taking Kant's 
Critique as the point of departure, whether one altogether 
adopts his conclusions or no. Some account must also be 
taken of the question as to what is implied in the possibility 
of natural science, before the parallel question as to morals 
can be dealt with. Until we have satisfied ourselves that the 


existence of a science of nature implies a principle in man 
which is not natural, the suggestion will always be returning 
that man, as a moral agent, is merely a very complicated 
natural phenomenon, to be treated as such by a purely 
natural science. 

73. The most striking, and at the same time most true 
and significant, expression of Kant's doctrine of knowledge is 
the statement that ' the understanding makes nature.' It is 
agreed that all which exists for us is what is given in experi- 
ence, but we speak of experience as depending on real things 
or objects or an order of nature. We distinguish the merely 
subjective or fanciful in our impressions from the objective, 
on the ground that the former consists in ideas that 'we 
make to ourselves,' or in arbitrary interpretations of ex- 
perience, as distinct from impressions which real objects 
make on us, or relations between impressions in the way of 
natural cause and effect. What then are real objects, order 
of nature, natural relations of cause and effect ? Macbeth's 
vision of a dagger is mere vision, merely subjective ; there 
is no real object causing it. What then would have been 
the real object if it had not been a mere vision ? There would 
have been a certain combination of moving particles, irri- 
tating the optic nerve in a certain way, and such as under 
certain conditions would produce many other sensible effects. 
But in saying this we have been describing a complex of 
relations, any one of which implies all the rest and derives 
its nature from the whole universe of possible experience. 
The real object is this complex of relations. There is no 
thing in which they reside or to which they belong, no object 
other than that which they constitute. Just as the real 
object, corresponding to or represented by any particular sen- 
sation, consists in the fact that this sensation is related in a 
particular way to other possible sensations, so the objective 
world as such or altogether consists in the series of such 
facts, in the system of relations between all possible sensa- 
tions, all possible data of experience. Such a system of 
relations implies a single relating principle (a principle of 
connection or synthesis or unity) on the part of the con- 
sciousness for which (or as an object to which) the relations 
exist ; a function exercised by consciousness, which yields 
this object consisting in a single system of knowable rela- 
tions. This principle or function Kant calls ( understanding*' 


or c synthetic unity of apperception, 5 which ' makes nature/ 
according to him, in the sense that it makes the data of 
sense into one system of related elements, the succession of 
feelings into an experience of objects. 

74. All the laws of nature, which we ascertain by experi- 
ment and observation, must, according to Eant's theory, be 
the work of understanding, connecting the data of sensi- 
bility according to certain forms ; but we may distinguish 
the particular laws of nature from those ways of connecting 
phenomena which are necessary to there being a nature at 
all, the laws ascertained by experiment and observation from 
principles presupposed by experiment and observation. The 
latter Kant called pure or a priori principles of the understand- 
ing (e.g. c all changes happen according to the law of the con- 
nection of cause and effect '). But though the understanding, 
according to Kant, c makes nature, 5 it makes it out of a mate- 
rial which it does not make ; ' macht zwar der Verstand die 
Natur, aber er Schafft sie nicht. 5 It makes it by connecting 
' intuitions,' data of sensibility, or phenomena given under 
the forms of intuition, space and time. Space and time are 
so far on a level with the ' forms of understanding 5 that they 
are (1) not sensations or results of sensitive experience, nor 
(2) conditions of 'things in themselves 5 as distinct from things 
of consciousness. They are conditions under which we are 
conscious of objects, but conscious in the sense of perceiv- 
ing as distinct from understanding. Kant habitually writes 
as if perception (' intuition 5 ) preceded understanding, as if 
the functions involved in the two operations were different. 
Indeed, if we take his statements as they stand, it would 
seem as if nature or knowledge or experience of objects 
implied three factors, (a) the presentation of mere sensations 
(' Empfindungen 5 ) from an unknown source, (b) the deter- 
mination of the matter so given by forms of the * intuiting 5 
consciousness, resulting in a mere manifold in space and 
time (' Anschauungen 5 or ' Erscheinungen 5 ), (c) the unifica- 
tion of this manifold by understanding. This separation 
of mere sensation from intuition or perception, and of percep- 
tion from intelligence, is generally admitted to be untenable. 
There seems to be a certain explanation of space and time in 
saying that they are conditions of a perceiving as distinct 
from an understanding consciousness, because, being familiar 
with the term perception, we suppose ourselves to know 


something about it, which gives meaning to the account of 
space and time as its necessary forms. But what, according 
to Kant, can we say of perception (intuition) but that it is 
consciousness under the form of space or time ? If we cannot 
say more than this, it is no explanation of space and time 
to say that they are forms of perception as distinct from 
understanding. In fact, all that can be said of the distinc- 
tion between intuition or perception and understanding, 
according to Kant, is that one and the same consciousness 
of an object is an act of intuition in respect of the manifold- 
ness of the object, i.e. in respect of its being an event follow- 
ing another event, followed by a third,, and so on, or in 
respect of its having a plurality of spatially distinct parts, 
and an act of understanding in respect of that unification of 
the manifold, without which there is no relation, no object de- 
termined by relation (properly, no object at all). But whether 
the distinction of sensibility with its forms from understand- 
ing with its forms be tenable or no, it remains true that the 
consciousness in which nature, or the c cosmos of experience,' 
arises, is subject to a certain condition in respect of the mode 
in which its material is presented to it, and that certain 
characteristics of nature and our knowledge of it arise out of 
this, or from the action of the unifying principle in relation 
to it. This is the condition which the Germans call ' Aus- 
sereinander-sein.' Intelligent consciousness is a perpetual 
process of unification of that which comes into consciousness 
in separation or bit by bit. Whether space and time are 
rightly treated as co-ordinate forms of this separateness 
may be doubted. It would rather seem that separateness in 
time simply was the distinguishing condition of what comes 
into intelligent consciousness so far as not yet affected by 
the unifying principle in consciousness, and that separate- 
ness in space, which implies coexistence and mutual limita- 
tion of the elements thus separate, is already a result of the 
action of the unifying principle upon data presented to it in 
succession, but which it combines, in negation of the succes- 
sion, as coexisting parts of a whole. 

75. Knowledge, then, according to Kant, is a process of 
unifying (of rendering into a perfectly interrelated whole) 
a material, which, owing to the conditions under which it 
appears (comes into consciousness), can never be completely 
unified. A complete or final determination of one 


menon by other phenomena can, from the nature of pheno- 
mena, never be reached, because every phenomenon implies 
another, before it in time or beside it in space. We ascertain, 
e.g., the nature of a certain event as determined by ante- 
cedent events (or, as Kant would say, the understanding 
gives a nature to a phenomenon, to something that enters 
into consciousness, by connecting it with what previously 
appears or enters into consciousness) ; but just because it is 
phenomena with which it is thus connected, phenomena 
subject to the form of time, i.e. which come into conscious- 
ness in succession, these in turn imply previous phenomena, 
and so ad infinitum. The same insoluble problem meets us 
in whatever way we seek for totality in our knowledge, or 
try to regard the world as a whole. We do not know it 
completely, it has not completeness, unless it is made up of 
simple parts, and is limited in space and time ; and each of 
these suppositions involves contradictions. The ultimate 
particles of which the world of experience is composed, if 
there are such, must occupy space, otherwise they would not 
belong to the world of experience at all ; but space consists 
of spaces, and what occupies space must occupy spaces, i.e. 
must contain a manifold the parts of which are external to 
each other; therefore the supposed ultimate particles are 
composite, i.e. are not ultimate. In the same way, as every 
time implies a time before it, every space a space beside it, 
a limit of the world in space and time is a contradiction. 

76. Clearly these insoluble problems, arising as they do 
out of the effort to know completely, originate in the same 
unifying principle in which, as determining the sensibility, 
all knowledge (and nature) according to Kant originates. 
Kant, however, is apt to call this principle by different 
names according as it acts, (a) in the way of connecting 
phenomena into a uniform cosmos, (b) in the way of suggest- 
ing the insoluble problems which arise out of this process 
of connecting phenomena. As acting in the former way he 
calls it understanding, in the latter way reason. Thus he 
treats understanding as the source of knowledge, reason as the 
source, not of knowledge, but of problems, or of ideas which 
cannot be realised in the relations of phenomena (and which 
are thus opposed to conceptions (' Begriffe '), since these always 
connect phenomena). These ideas are of the absolute or 
complete or unconditional in various forms. They have, 


according to Kant, a regulative, though not a constitutive, 
use in knowledge, i.e. they do not serve to constitute or 
present to us an object, which can be known as absolute, 
complete, or unconditional, for to knowledge is necessary the 
presentation of an object as related to other objects in space 
and time in a manner incompatible with its being complete 
or absolute ; but they regulate or determine that activity of 
consciousness which, in relation to the sensibility, under the 
designation ' understanding,' results in knowledge. The idea 
of there being an unconditioned is the source of the quest after 
a totality of conditions which forms the process of knowledge. 
Precisely the same principle in consciousness, therefore, 
which as reason is the source of the ideas to which no phe- 
nomena or relations of phenomena correspond, and of the 
problems to which knowledge can yield no solution, renders 
knowledge and nature possible. The ideas and problems of 
reason, in short, are simply the beginning and end of know- 
ledge ; the beginning, in the sense that in the consciousness 
of an unconditioned, which is reason's consciousness of itself, 
originates that search for a complete sum of conditions 
which constitutes the process of knowledge ; the end, in the 
sense that the search for such a sum within the data of sen- 
sibility issues for the reasons given in insoluble problems. 

77. Thus when it is said that according to Kant ' know- 
ledge is merely of phenomena,' or of nature as the complex 
of (the system of relations of) phenomena, we must get rid 
of the notion that the object-matter to which knowledge is 
thus said to be confined is something apart from and inde- 
pendent of knowledge or the intelligence which knows. If 
we speak of phenomena, or the objects of knowledge, putting 
limits upon knowledge, we must remember that these limits 
are part and parcel of knowledge itself, that they are simply 
incidents of the knowing process ; in other words, that the 
nature, to which the operations of intelligence are confined, 
is itself the work of intelligence, and the insoluble problems 
which nature presents to the understanding are the under- 
standing's own making. It is the intelligent synthesis of 
phenomena which yields the insoluble problem of reaching 
completeness of synthesis ; e.g. it is through the holding to- 
gether by intelligence of times, the addition of spaces, that 
there arises the infinite series of time and space which seems 
to baffle intelligence. 


We must remember, further, that, if knowledge is only 
of phenomena, or, more properly, only consists in the 
establishment of relations between data of sensibility, there 
is at least such a thing as a reflective analysis, whether 
we call it ' knowledge ' or no, of what is involved in this 
process of knowledge (an analysis represented by Kant's 
own treatise), by which it is ascertained that the existence 
of a knowable nature implies that of a principle of union 
which is not itself part of the knowable nature, not one or 
any number of the relations which constitute it ; an uncon- 
ditioned, in relation to which alone the mutual conditioning 
of phenomena is possible ; a consciousness of laws of nature, 
or rather a principle of consciousness which, in relation to 
sensibility, yields laws of nature, which is not itself subject 
to those laws of nature. 

78. The question, then, arises (1) whether this principle 
of consciousness, called from different points of view the 
unconditioned, reason, and understanding, has any other 
function than that which is exercised in the constitution of 
intelligent experience, and of nature as = the cosmos of ex- 
perience ; and (2) if it has, whether there is a philosophy of 
it, and, if there is, how it is related to the knowledge of 
nature. Kant in effect holds that it has another function, 
one distinguished from that which constitutes knowledge, as 
being exercised in relation to desire. 

1 Desire is consciousness of a wanted object. As certain 
conditions of sensuous excitement, in relation to a self- 
conscious subject distinguishing itself from its conditions, 
become sensible objects, so a condition of want or appetite, 
in relation to such a subject, becomes consciousness of a 
wanted object. The sensible object is something which is ; 
the wanted object (the filling of the want) is something 
which is to be (has yet to be brought into existence). In 
this lies the distinction between ' sein ' and ' sollen ' in the 
most elementary form. As intelligent experience, and with 
it nature and knowledge, result from the presentation of 
sensible objects and their connection in one universe (a con- 
nection which results from that same relation to a self-con- 
scious subject which is the condition of their presentation 
as objects), so practice results from the presentation of 
wanted objects, objects to be brought into existence. But 

1 [Cf. Prolegomena to Ethics, §§ 86-87.] 


whereas in knowledge the sensible object carries its reality 
with it (in being presented at all it is presented as real), in 
practice the wanted object is one to which reality has yet to 
be given. (If I want a mutton chop or a picture of Turner, 
the chop and the picture are no doubt in existence, but the 
object is the filling of my want by eating the chop or acquisi- 
tion of the picture, and that is an object which has yet to be 
realised.) Thus the world of practice depends on man in 
quite a different sense from that in which nature does. We 
commonly speak of nature as wholly independent of man. 
This is not true in the sense that there could be nature (the 
nature that we know) without intelligent consciousness ; but 
it is true in the sense that, given the consciousness of 
sensible objects, it does not depend on any exercise of our 
powers whether they shall become real or no; they are 
already real. On the other hand, in the world of practice 
consciousness of an object is prior to its reality, and it 
depends on a certain exercise of our power, determined by 
that consciousness, whether the object shall become real or 
no. The question then arises, how this consciousness of 
objects, which in the world of action precedes and condi- 
tions their existence, is itself determined. Prima facie it 
distinguishes the series of moral actions from any series 
of natural events; since in the latter a preceding con- 
sciousness of the event is not a condition of the event's 

79. l A ' naturalist' will say that this is an arbitrary 
limitation of the natural ; that a determination of conscious- 
ness having natural antecedents is a natural event as much as 
any other ; and thus that the fact that the direction of our 
powers in moral action to realise the object of desire is con- 
ditioned by consciousness of the object does not denaturalise 
moral action ; that, if it did, on the same principle we must 
reckon the actions of animals, which seem to be conditioned 
by the consciousness of objects wanted, other than natural. 
The question turns upon the action of self -consciousness in 
the determination of the action which we say is not strictly 
natural. No one pretends that an appetite or want is other 
than a strictly natural event, or that its effect in the way of 
an instinctive action directed to satisfy the want is so either. 
But it is contended that the consciousness of a wanted 

' [Cf. Prolegomena to Ethics, §§88 ff.] 


object, the presentation of satisfaction of want as an object, 
is quite different from mere want ; that it implies self-con- 
sciousness, consciousness of itself on the part of a subject of 
the want as such a subject ; that thus a motive is constituted 
quite different from mere want, and that it is this (conscious- 
ness of self-satisfaction to be attained) which in all cases 
determines action properly called moral. Upon the motive 
constituted by the presence to self-consciousness (the taking 
up into self-consciousness) of a simple animal want, there 
supervene all sorts of other motives, as what we call human 
nature developes (a development arising from that widening 
consciousness of a world which self-consciousness renders 
possible) ; but the common form of all motives is the con- 
sciousness or presentation of a self-satisfaction to be attained, 
however widely the conditions of the self-satisfaction vary. 
Is then this self- consciousness (which may equally be called 
the consciousness of ends, and out of which arises, as will be 
explained, the consciousness of practical laws) itself natural ? 
Nature = the system of sensible events or objects as inter- 
related. That is natural which is either a sensible event or 
object (as so related), or a relation between such events or 
objects. The self-conscious principle, implied in the pre- 
sentation of self-satisfaction as an object, is not such an 
event, object, or relation. Rather it is identical with the 
principle in virtue of which there is for us a nature. 

80. But it may be objected, 6 If you say that understand- 
ing, which you identify with the principle of self-consciousness, 
makes nature, how can you oppose moral action to natural 
phenomena on the ground that the former are, while the 
latter are not, determined by self -consciousness ? The 
animal want, as an event or phenomenon in the order of 
nature, is already, according to your showing, something 
determined by self-consciousness. What other determina- 
tion by self-consciousness is it that makes it cease to be a 
mere animal want, and yields instead a moral motive ? ' 

1 The statement that 6 the understanding makes nature ' 
may be understood in two ways. Is it meant that only 
through understanding there is such a thing as nature at 
all? or that only through understanding, as a principle in 
us, there is for us a nature, i.e. the data of our experience 
are so connected as to yield the consciousness of there being 

1 [Cf. Prolegomena to Ethics, § 19, and elsewhere.] 


a nature ? Primarily, at any rate, it is in the latter sense 
that the statement must be taken. Nature, as a determinate 
order of phsenomena, exists independently of the conception 
of nature as gradually formed by any of us. It is difficult, 
indeed, if we think the matter out, to come to any other 
conclusion than that the principle in us, which, through its 
equal presence to and distinction from all the data of our 
sensibility, connects them all into a related series of changes, 
is identical, under however distinct a form, with the prin- 
ciple through which nature exists as such a series, before and 
independently of the experience of any of us. When it is 
said, then, that understanding makes nature, either ' un- 
derstanding ' must be taken as other than a principle or 
function of the individual man, or, if taken as belonging to 
any individual man, to 4 make nature ' must mean to cause 
there to be a nature for that individual man. On the other 
hand, when it is said that self-consciousness makes moral 
action, it is meant that it is through self-consciousness on 
the part of the individual man that there is such a thing as 
moral action, whereas it is not through understanding, as 
on the part of the individual man, that there is such a thing 
as nature. 

81. To return, then, to the airopla put above. The animal 
want, both as in itself (a) an event in the order of nature, and 
(b) as known by us to be so, is in a certain sense determined 
by relation to a self-conscious subject ; but, so long as it 
remains a mere animal want, so long as it is an event in the 
order of nature or knowable phsenomena properly so called, 
it is not determined by a subject conscious of itself as 
affected by and giving a character to the want. When it is 
said that any animal want, as belonging to or conditioned 
by an order of nature, is determined by relation to a self- 
conscious subject, it is meant that a system of relations 
between facts, such as we understand by c nature,' only exists 
for (in virtue of there being) a single subject, distinguishing 
itself from the facts, but so present to them all as to hold 
them in relation. The state of my digestion, then, as at 
any time it happens to be, like any other fact in the order 
of nature, implies the subject presupposed in the possibility 
of such an order. But this subject is obviously not itself 
the subject of any or all the particular states, which through 
relation to it are related to each other ; avdy/cr) dfuyr) shcu, 


toairep (prj&lv 9 Ava£ay6pa? 9 tva /cpary. 1 If itself conditioned by 
any of the changes which it unites in one system, it could not 
so unite them, (b) Let us next consider the animal want as 
known by us to be an event in nature. Such knowledge of 
it is determined by self- consciousness, in so far as the con- 
sciousness of self is the necessary correlative to the conscious- 
ness of there being nature ; but in being known as an event in 
nature, the want is not affected by, and does not affect, the 
self-conscious subject of which the action is necessary to its 
being so known. For the consciousness of the animal ex- 
periencing it there may very well be no nature and no know- 
ledge. A man indeed may have animal wants and at the 
same time understand them (recognise them as belonging 
to an order of nature) ; but the condition of his so under- 
standing them is that he hold himself aloof from them, that 
he do not introduce himself, the self-conscious, understanding, 
subject, as a qualifying element into them. Just in so far 
as they are taken up into his personal consciousness, just 
so far as he is conscious of himself as affecting or affected 
by them, there supervenes upon the animal want a new 
experience, which is not properly a natural phenomenon, 
or knowable as such. When the poet, e.g., introduces his 
personal consciousness into the natural phenomena which 
he contemplates, when he modifies them by the special 
relation which he establishes between himself and them, the 
result is something which is not a natural phenomenon or 
knowable as a matter of fact, ' not fact but poetry.' In the 
same way, when the satisfaction of a purely animal want, or 
a pleasure having strictly natural conditions, is taken up 
into self-consciousness (when a self-conscious subject makes 
it its own), it ceases to be a naturaL phenomenon; it 
becomes an interest having a moral as opposed to a natural 

82. Thus self-consciousness is an agent within the series of 
moral actions as it is not within the series of natural events. 
As implied in understanding (' unity of apperception ') it 
is necessary to our being conscious of a nature, but it does 
not directly condition any of the phenomena which in virtue 
of it we regard as conditioned by each other. On the con- 
trary, it is essential to their being understood as belonging 
to or forming a nature, that they be not qualified by the 

1 [Aristotle, De Anima, iii. 4.] 


personal consciousness of the man who understands them. 
If we hold, as we are entitled to do, that a self-conscious 
principle, from which that in each of us is derived, is the 
condition of there being a nature at all, still this is not an 
agent within the related series which it renders possible. 
The conditions of any natural event or existence are exclu- 
sively to be found in other natural events or existences, 
not in that which is the condition of their conditioning one 
another. On the other hand, in the moral world, reason as 
the self-consciousness of the individual man directly affects 
and gives its distinctive character to each action that goes 
to constitute this world. Whereas in the order of nature 
events happen in a determinate series, whether the under- 
standing of the individual connects them for him in such 
a series or no, in the moral world it depends on the pre- 
sentation by the individual of an object to himself, as one 
which will yield him personal satisfaction, whether an action 
is done or no. In other words, it is characteristic of moral 
action to he free, in a way in which no event in nature is 
free, and which differences the philosophy of moral action 
from any natural science. 

83. The assertion that human action is free in this sense 
is quite compatible with the admission that every act is deter- 
mined by the strongest motive. So far from free action being 
unmotived, it is rather determination by motives, properly 
understood, that constitutes freedom. A motive always implies 
consciousness by the individual of his own good as his end. 
Whatever good he presents to himself, the most ' altruistic ' 
good possible, is necessarily conceived as related to himself, 
and is determined as good by that relation. c That is good ' 
= 'that satisfies me.' An agent determined by a motive, 
then, is determined by himself, by that consciousness of 
himself as the absolute or unconditional end which makes 
the motive, and is so far free. There is no doubt another and 
higher sense of moral freedom than that on which we are 
now dwelling, and which is equally characteristic of the worst 
act and the best. What we are here describing (to adopt a 
distinction used by some German writers) may be called 
formal as distinct from real freedom. The real or higher 
freedom is only attained so far as the ends in which self- 
satisfaction is sought are such as can really satisfy. 

84. Reverting, then, to Kant's statement, ' everything in 


nature works according to laws ; the distinction of a rational 
being is the faculty of acting according to the consciousness 
of laws/ we may adopt it if by 'faculty ' we understandpos- 
sibility (as we always should) , and recognise this possibility as 
lying in the consciousness of self and of ends relative to that 
self. Action according to the consciousness of laws clearly 
presupposes the consciousness of ends to be attained by con- 
formity to these laws. The latter consciousness may deter- 
mine action without the former having been developed, and, 
when it has been developed, in spite of it. Perhaps in 
certain characters the conception of law may determine 
action apart from the consciousness of any end other than 
the law, but then the fulfilment of law is itself presented as 
an end. For Kant's formula, at the cost of spoiling its 
antithetical compactness, we may substitute the following : 
■ Everything in nature works so as to yield certain results 
according to law ; the distinction of a rational, or free, 
being, is that he acts, not so as to yield certain results, but 
from consciousness of ends in attaining which he may satisfy 
himself, out of which arises the consciousness of laws accord- 
ing to which they are to be attained.' 

[A discussion of human freedom follows, ' on the lines of 
Kant, who first clearly brought out the difficulties of the 
question, and furnished elements (though no more) for a 
true answer.' . The substance of this discussion is embodied 
in the Prolegomena to Ethics, and in the articles on H. 
Spencer and G. A. Lewes in vol. i. It concludes as follows : 
' Thus in a certain sense, though one different from Kant's, 
we may adopt the distinction between the ( empirical ' and 
' intelligible ' characters. A moral action has one character as 
a. natural phenomenon, related in the way of cause and effect 
to other natural phenomena ; another character as expressing 
the desire or emotion or thought of a self-conscious man. 
According to the latter character it is not a phenomenon at 
all. The order in which it has its determining place is not 
an order of such phenomena. In respect of its intelligible 
character action is free, not so in respect of its empirical 
character. It is in another sense, however, that Kant dis- 
tinguishes the two characters. For he would apparently 
reckon the relation of any action to a man's mental his- 
tory (all, in fact, by which we account for it), as belonging 


to its empirical character, whereas, according to the distinc- 
tion which it is here sought to maintain, the conditions 
which form the true explanation of the act (the other deter- 
minations of consciousness, as opposed to natural events, to 
which it is related) would be assigned to the intelligible 

Then follows an examination of Kant's doctrine of * free 
causality,' as developed in the Critique of Pure Reason, in 
connection with the third i Antinomy of Pure Reason.' This 
forms the following section, L.] 

VOL. II. h 



85. Accokding to Kant, both the thesis and the antithesis 
of the third antinomy are true in different relations. This 
applies to the antinomy both in its cosmological and its 
ethical bearing. First, cosmologically. There is no free 
causality within the world of phenomena, but free or un- 
conditioned causality is the condition of there being such 
a world. We might thus combine the thesis and anti- 
thesis by saying, ' Everything in the world happens accord- 
ing to laws of nature, but a causality of freedom is necessary 
to account for there being laws of nature.' The antinomy 
itself is evidence of there being an unconditioned causality, 
of which the effect is that events form u series each deter- 
mined by its antecedent. It is the fact of their forming 
such a series that leads to one of the antithetical proposi- 
tions (that which denies freedom), but the fact is itself an 
effect of an e unconditioned causality.' According to Kant's 
way of putting it, 1 if the phenomena which form the series 
were ' things in themselves,' if their nature belonged to them 
in their own right, independently of anything else which is 
not phenomenal, this would not be so. Each phenomenon 
would be (as it is) determined by a preceding phenomenon, 
this by another, and so on, and there could be no ques- 
tion as to any other mode of determination. In fact, they 
only form a series through relation to the unconditioned. 
Phenomena = sensible events. If there were nothing besides 
these, as the order of their succession in time does not admit 
of an absolute beginning, there could be no such thing as 
free causality. But there must be something besides them. 
A determinate succession of presentations to sense implies 
something which determines them, and which cannot be any 

1 Kritik der reinen Vernunft, pp. 372-373 ; pp. 332-333, TV. 


one or any number of them ; a subject present to them all 
which renders the connection of any one with all possible. 
When Kant says that phenomena are not c absolutely real ' 
but ' mere representations, 5 he means, as he explains else- 
where, 1 that they f cannot exist out of and apart from the 
mind,' or * apart from and independently of experience,' 
which experience is rendered possible by the unity of the 
transcendental subject. Again, ' they are mere representa- 
tions, receiving from perceptions alone significance and 
relation to a real object, under the condition that this or 
that perception — indicating an object — is in complete con- 
nection with all others in accordance with the rules of 
the unity of experience.' 2 lb would seem, then, that, accord- 
ing to Kant, the non-phsenomenal ground of phenomena, 
as 'connected with each other according to empirical 
laws,' 3 must be what he elsewhere calls ' the transcendental 
ego,' or ' subject,' as the source of the unity of experience. It 
is true that in the passage referred to 4 he speaks of it as 
' transcendental object,' which he elsewhere calls 5 the ' non- 
sensuous cause of phsenornena,' the ' mental correlate to sen- 
sibility considered as receptivity.' But from the same passage 
it appears that, just as any empirical object means a per- 
ception as connected (or the connection of a perception) 
' with all others in accordance with the rules of the unity of 
experience,' so the ' transcendental object ' means that which 
connects all phsenomena, viz. the ' transcendental ego ' as 
the source of the unity of apperception, considered specifically 
in opposition to the receptivity of sense as an active cause. 
Thus we find that the unconditioned condition of the order 
of nature, as Kant understood nature (i.e. as the connection 
or unity of experience), is what he has previously called the 
6 transcendental ego.' This is that of which all sensible objects 
or phsenomena are effects, in the sense that objects are percep- 
tions as connected, and this is the source of connection. It is 
not, however, in Kant's language, an object of possible know- 
ledge or experience (though the source of all knowledge), 
because not a phsenomenon. All knowledge consists in con- 
necting phsenomena with phsenomena, and can never admit 
within its world an uncaused cause. So far, in regard to the 

1 E.g. Krit. d. r. V. p. 347 ; p. 307, Tr. * Ibid. p. 374 ; p. 333, Tr. 

2 Ibid. p. 349; p. 309, Tr. * Ibid, p. 349 ; p. 309, Tr. 
8 Ibid. p. 373 ; p. 332, Tr. 

h 2 


cosmological idea of freedom, there is no particular difficulty. 
The question as to an absolute beginning of nature is set aside 
by the reflection that to seek for it is to seek for that in 
time which is the condition of there being time. It is the 
essence of time that it can have no beginning, but if time is 
the sequence of representations, this merely means that there 
is no beginning of consciousness. But though there is no 
absolute beginning of nature, there is an absolute or free 
cause of it. All phenomena as known are connected in the 
way of cause and effect with other phenomena, but in respect 
of this connection, in respect of their knowability, are effects 
of the intelligible cause in virtue of which there is such con- 

86. The possibility of ethical freedom is another matter, 
for this implies that an object (this or that man, this or that 
state of mind), which we know as a phenomenon, and which 
as a phenomenon is not a free cause but conditioned by other 
phenomena, is yet also not a phenomenon, has an intelli- 
gible as well as an empirical character, and, as such, is a free 
cause. For the fact, however it is to be explained, that man 
is a free cause, Kant appeals to the existence and effect of 
the conception s I ought. 5 This is not a conception of any 
natural or phenomenal object. ' The question, What ought 
to happen in the sphere of nature ? is just as absurd as the 
question, What ought to be the properties of a circle ? ' ■ The 
action of the motive ? I ought ' is not the action of a pheno- 
menon, nor itself the result of a phenomenon. Man, there- 
fore, as capable of such motive, is other than a phenomenal 
cause. 2 But though other than a phenomenal cause, man is 
also an object of the senses. His acts are phenomena, 
connected according to laws of empirical cause and effect 
with other phenomena. Kant expresses this by saying that 

1 Kant quite recognises that man is a cause of phenomena ; in respect of 

' a purely intelligible object,' as opposed reason he is. 

to a ' phaenomenon,' in respect of under- 2 Though Kant appeals to the capa- 

standing also, the source of pure ' con- city for the motive ' I ought ' as evi- 

ceptions,' in distinction from pure 'ideas' dence that man is a free cause, he does 

(Krit. d. r. V. p. 379 ; p. 338, Tr.). Pure not (on the whole) mean that only ac- 

conceptions are those by which we con- tions determined by this motive are free, 

nect phenomena so as to form a nature, 'Keason is the permanent condition of 

a cosmos of experience. Understanding, all actions of the human will ' (p. 382 ; p. 

the faculty of these conceptions, accord- 342, Tr.). It is the condition, we must 

ing to Kant, is merely reason in its suppose him to mean, even of actions that 

application to phenomena. In respect proceed from what he calls a patholo- 

of understanding, however, man is not gically affected will (p. 371; p. 331, Tr.). 


the causality of man is twofold, intelligible and sensible ; 1 
and again, that man has an intelligible and an empirical cha- 
racter. (Character = the law of his causality.) The ' intelligi- 
ble character ' represents the relation of man as a free cause 
to his acts, the ' empirical character ' the relation of his 
actions, as phenomena, to each other and to all other phe- 
nomena, a relation in virtue of which they can be ' deduced 
from these other phenomena as conditions, and thus, in 
connection with these, constitute a series in the order of 
nature.' 2 The ' empirical character ' alone is, in Kant's 
sense, knowable, because to constitute a knowable object 
there must always be a phenomenon or object of possible 
intuition (perceivable), as well as a conception which con- 
nects and relates it with other objects. If we want to know, 
or account for, an action done by any one, we must consider 
it as a fact in his ' empirical character,' ' to be accounted for 
by reference to preceding phenomena.' 3 Yet, according to 
Kant, ' this empirical character is itself determined by the 
intelligible character ' 4 in which ' nothing happens,' f which 
knows no before and after,' but of which ' every action, irre- 
spective of the time-relation in which it stands with other 
phenomena, is the immediate effect.' 5 This doctrine is 
generally pronounced very unsatisfactory. ' It seems (it may 
be said) to imply that the same man has two characters, one 
which can be known, another which is unknown, and that 
the unknown is the cause of the known. Either it means 
nothing, or it violates Kant's own doctrine that a free cause 
can never be admitted within the series of phenomena.' 

87. The objection partly arises from Kant's use of the 
term e character,' which, however, he explains. 6 He did not 
mean that the same man had two characters, but that one 
and the same character (in the ordinary sense of the word), 
i.e. series of acts (inner as well as outer), was related at once 
to an intelligible cause consisting in reason, and to a series 
of empirical causes consisting of other phenomena. ' Well,' 
it will be said, ' but the question is about the compatibility 
of these relations ; how the same moral act can be caused in 
these different ways, arid this Kant does not explain.' The 
true explanation is that these c empirical causes ' in turn, as 

1 Krit. d. r. V. p. 374 ; p. 333, Tr. * Ibid. p. 381 ; p. 340, Tr. 

* Ibid. p. 374 ; p. 334, Tr. s Ibid. p. 383 ; p. 342, Tr. 

* Ibid. p. 375 ; p. 334, Tr. s Ibid. p. 374 5 ; p. 334, Tr. 


causes of a human character or of a moral act, are what they 
are through relation to the intelligible cause. At any rate, 
Kant points out that each member in a series of phenomenal 
causes and effects may be determined by relation to a cause 
which is not phenomenal. It would have been simpler if he 
had said at once that a human action is not, properly speak- 
ing, a phenomenon (not a phenomenon in the sense in which 
a phenomenon derives its nature from relations in the way 
of time and space to other phenomena), though it involves a 
phenomenon as its expression, and that it could not be 
accounted for by reference to other phenomena merely as 
phenomena. No doubt it is accounted for by reference to 
preceding actions, but these really account for it just so far 
as they are other than mere phenomena, just so far as each 
is an expression of some mode or other of that self-con- 
sciousness, which is the condition of there being natural 
phenomena for us, but is not itself a natural phenomenon. 

'Is it not possible, 5 asks Kant, 'that, although every effect 
in the phenomenal world must be connected with an empiri- 
cal cause, according to the universal law of nature, this em- 
pirical causality may be itself the effect of a non-empirical 
and intelligible causality, its connection with natural causes 
remaining nevertheless intact ? ' l In that part of the phe- 
nomenal world which is the expression of moral action, this 
is not only possible but real. The ' determinists ' are quite 
right in saying that what a man is and does at any time is 
the result of what he has previously been and done ; but 
what he has previously been and done, though in respect 
of its physical expression belonging to the phenomenal world 
and connected in the way of antecedence and consequence 
with mere natural phenomena, has taken that specific 
character in virtue of which it determines what the man 
now is and does, from relation to reason as a c non-empirical 
and intelligible causality.' His previous character has been 
a state of self-consciousness. It is the result of a process 
in which the presentation of a self to be satisfied, of an end 
to be attained for the satisfaction of self, has been the 
dominant and determining agent. His previous actions 
have been the expression of desires or emotions belonging to 
a universe of consciousness, into which natural events, as 
such, do not enter (only enter as transformed by the personal 
I Krit. d. r. V., p. 377 ; p. 336, Tr. 


consciousness of the individual). Thus, though the connec- 
tion of human actions with each other is as regular and 
admits of being as definitely known (though the knowledge is 
more difficult to arrive at) as that of natural phenomena, 
the things connected are different, because in the latter case 
the unconditioned (self-consciousness) does not qualify any 
one of the things connected, except as rendering possible its 
connection with other things in experience; in the former 
case it qualifies, as consciousness of self or law, each of the 
things connected. 

Kant, however, even while saying that c every action . . . 
is the immediate effect of the intelligible character of pure 
reason ' l seems to consider that human actions, as alone they 
can be known, or as objects of speculative reason when we 
try to ' explain their origin,' 2 form a mere series of natural 
events. ' So far as relates to this empirical character, there- 
fore, there can be no freedom, and it is only in the light of 
this character that we can consider the human will, when we 
confine ourselves to simple observation, and, as in the case of 
anthropology, institute a physiological investigation of the 
motive causes of human actions.' 3 And again, ' the natural 
law, that everything which happens must have a cause — that 
the causality of this cause, that is, the action of the cause . . . 
must have itself a phenomenal cause, and consequently that 
all events are empirically determined in an order of nature — 
this law, I say, which lies at the foundation of the possibility 
of experience, and of a connected system of phenomena or 
nature, is a law of the understanding, from which no depar- 
ture and to which no exception can be admitted.' 4 In fact, 
however, if his account of moral action be true, viz. that it 
is an ' effect of the causality of reason,' in explaining it or 
trying to know it as a mere natural phenomenon determined 
by a preceding natural phenomenon, we are explaining 
and knowing it wrongly. It may be true that mere ' obser- 
vation ' of the actions of another could not enable us to know 
them in any other way, because this is understood to exclude 
the explanation of such actions in the light of that self-know- 
ledge which man has through ' pure apperception ' ; 5 but can 
we only know what we can ' observe ' ? According to Kant's 
use of the term ' knowledge,' that is so. Knowledge with 

1 Krit. d. r. V. p. 383 ; p. 342, TV. * Ibid. p. 376 ; p. 335-6, TV. 

2 Ibid. p. 381 ; p. 340, TV. s Ibid. s Ibid. p. 379 ; p. 338, TV. 


him is indeed of relations, but of relations as between phe- 
nomena or sensible objects. The relation, then, between 
action and reason, as a ' free cause,' would not be matter of 
knowledge. Kant thinks indeed that the relation between 
actions, since actions are phenomena, is matter of knowledge, 
but it is not easy to see how he can escape the retort, that 
in knowing them as phenomena we are knowing them as 
being that which, according to his doctrine, they really are 
not. 1 

88. To sum up the criticism of Kant : the worst of his 
account of the intelligible and empirical character is that 
it seems to keep the c free cause ' outside the phenomena of 
human action. Either (1) it stands, according to him, to 
these phenomena merely in the same relation in which it 
stands to phenomena of nature ; or (2), so far as he admits 
it to stand in any other relation to them, it is at the cost (a) 
of contradicting his principle that a free cause may not be 
admitted within the series of phenomena (a contradiction only 
to be avoided by the admission that human actions are other 
than phenomena), (6) of making the series of human actions, 
as they truly are, something different from the same as know- 
able. (1) Because free (unconditioned) causality, according 
to his view, conditions all empirical causality in nature. 
The phenomenal cause of a phenomenon, the sensible event 
which invariably precedes another, is only its cause in virtue 
(as natural philosophers would say) of the co-operation of all 
the conditions of the world, of all that is or has been. This 
is not a phenomenon. If it is the sum of all conditions, it 
is itself the unconditioned. In the search for phenomenal 
causes, or antecedents, all science presupposes it (under the 
designation ' nature ' or ' order of nature '), though it never 
finds it, never admits it within the series of phenomena. 
According to Kant's view of nature, this conception of an 

1 It is idle to say, on the one hand, which has become his. This relation 
that a moral action is really other than forms the reality of the act, as a moral 
a phsenomenon, connected in the way of act. Yet this is what in many passages 
natural cause and effect with other Kant seems to understand by ' empirical 
phsenomena, and that thus it is free ; character.' It would be a different 
and, on the other, to say that in respect matter if ' empirical character ' meant 
of its ' empirical character ' it is a phse- merely what is strictly natural or phse- 
nomenon, so connected, if by its ' em- nomenal in the act, i.e. the relation of 
pirical character '. we understand its its physical expression to other facts of 
relation to the motives of the agent, nature. [Extract from a later lecture 
and through these to his past history in the same course.] 
and to that consciousness of others 


unconditioned causality (the conception of it as = the sum 
of possible conditions, or nature as a whole, which, he 
would say, leads to antinomies, since there can be no such 
sum, nor can nature form a whole), is replaced by the con- 
ception of it as the one subject which is the condition of the 
unity of experience. But, either way, empirical causality, 
as = the determinate sequence and antecedence of phenomena, 
presupposes an intelligible and unconditioned causality. And 
it is in no other sense, according to what Kant says of 
the relation between the intelligible and empirical charac- 
ters on pp. 377 and 381, (336, 340, Tr.) that the moral life 
implies an unconditioned causality. (2) On the other hand, 
as regards the second alternative, his whole ethical doc- 
trine turns on the supposition that reason, as free causality, 
is constitutive of the several acts of that life as it is not 
of 'natural phenomena ' ; but it does not appear how it can 
be so compatibly with what he says of these actions, either 
as a series of phenomena within which a free cause is inad- 
missible, or as an object of knowledge. 

We have, then, to deal with the following questions : (1) 
Is reason thus constitutive of the several acts of the moral life 
as it is not of natural phenomena ? and if it is, what is the true 
way of understanding the freedom of man ? (2) Adopting 
this view of moral action as that in which reason, as free 
cause, is a factor, how are we to reconcile it with the phe- 
nomenal character of such action, its character as one of a 
series in which free cause is inadmissible ? Or, if we deny 
that moral action is properly a phenomenon, how are we 
to explain its relation to natural phenomena ? (3) If we 
admit free causation within the series of human action, how 
is knowledge of them possible ? 



89. The problem with which Kant's moral theory com- 
pels him to deal, as he understands it, is this, ' whether 
freedom and natural necessity can exist without opposition 
(contradiction ?) in the same action.' It compels him to deal 
with this problem because his analysis of moral consciousness 
brings him to the conclusion that the only morally good 
action is one determined by the agent's conception of himself 
as, in respect of his rational nature, an absolute end. An 
action, a determination of will, so determined is ' independent 
of causes foreign to the will or agent,' and is in that negative 
sense free, as well as in the positive sense of being deter- 
mined by a law which the agent imposes on himself. Free- 
dom, then, in this sense of a determination by a reason 
which constitutes the agent's conception of himself as an 
absolute end or giver of law, must belong to an act if it is to 
be morally good. 

On the other hand, natural necessity * exists in the act,' 

(1) in so far as it is an object of observation, or of knowledge 
in that restricted sense of ( knowledge ' in which it is con- 
stituted by the connection of phenomena with phamomena 
according to the categories or formal conditions of experience ; 

(2) in so far as the action stands in that relation to desires 
and inclinations, not themselves determined by or a product 
of reason, which is implied in the fact that the rational con- 
ception of self as absolute end or giver of universal law acts 
as an imperative. It would not act as an imperative, there 
would be no distinction between ( I ought ' and ( I would,' 
unless in every action we were affected by desires, to which 
the rational conception is antagonistic, and which are them- 
selves of natural origin. The relation of the action to these, 

1 Krit. d. r. V. p. 385 ; p. 344, Tr. 


whether they completely determine it or are controlled by 
the purely rational conception, is one of natural necessity, 
not of freedom. 1 

90. How, then, is it explicable that these antagonistic 
attributes of freedom and natural necessity should belong to 
the same act? Contradictory attributes cannot belong to 
the same subject in the same relation. Kant's explanation, 
then, is that we ' think of man in a different sense and rela- 
tion when we call him free, and when we regard him as 
subject to laws of nature as being part and parcel of nature.' 2 
Very well, the reader says, but the question is not of the 
different ways in which we may think of man or of ourselves, 
but of the way in which he really exists. Is it possible that he 
should really exist in one relation as free, in the sense of being 
determined (actually or potentially) by a causality not foreign 
to him and not natural (not itself determined by any natural 
antecedents), viz. that of reason, and in another relation as 
determined by a causality that is foreign and natural and 
determined by natural antecedents, viz. that of desires and 
inclinations? If it is possible, it would seem that it can 
only be so in one of two ways. Either (a) the man is deter- 
mined by pure reason, and is thus free, in respect of some of 
his acts, determined by natural causes, and thus not free, in 
respect of others of his acts ; or (b) if it is one and the same 
act that is determined in these opposite ways, it must be a 
case of joint determination ; reason and the natural causes 
must co-operate in determining the act. 

91. Neither of these alternative views, however, seems to 
be either admissible in itself or admitted by Kant, (a) is not 
admitted by Kant, because, according to some passages, he 
regards a man as free in respect of vicious acts, which 
excludes the first alternative ; 3 though, on the other hand, 
according to his identification of freedom with autonomy 
and his explanations of autonomy (he does not regard the 
'cool villain' as autonomous), it follows that the vicious act is 
not free. (Was Kant's view that, though the vicious act is 
not free, yet a man is free to do or not to do it ; that he 
freely submits to the loss of freedom, the bondage of 
heteronomy ? In such a view, freedom is used in two senses. 
The submission could not be said to be rationally determined 

1 Grundlegung, &c, pp. 301-2; pp. 105-106, Tr. 
2 Ibid. p. 304 ; p. 110, Tr. 8 Krit. d. r. V. pp. 383-4 ; p. 343 TV. 


in Kant's sense ; therefore the man does not freely submit in 
this sense of freedom, but in the sense of having power to 
do or not to do ; whereas the loss of freedom to which the 
vicious man submits is the loss of it in the sense of rational 
determination. I think there is this double meaning of 
freedom in Kant.) As to alternative (h), it is excluded (1) 
by the general tenor of Kant's doctrine, which puts deter- 
mination by reason in direct antithesis to any determination 
by desire (except such as is excited by the pure contempla- 
tion of the moral law) ; (2) by his words, c This causality of 
reason we do not regard as a co-operating agency, but as 
complete in itself.' I 

92. Returning, then, to Kant's statement, ' we think of a 
man in a different sense,' &c, our conclusion must be that of 
these different senses or relations in which we may think of 
a man, it can only be in one or the other that the real man 
(as distinct from some element or factor of human nature, 
which we may consider in abstraction, but which is not the 
man) can exist and act. Now, according to Kant, a man ' can 
only act under the idea of freedom,' i.e. under the idea that 
he himself determines himself or his action (which is a dif- 
ferent thing from the consciousness of power to do or not to 
do), and c just for that reason is in a practical point of view 
really free.' This, I think, is quite true, and I should add 
that this ' practical point of view ' is the only point from 
which the man whom we contemplate is the real man. The 
man whom we contemplate from that point of view from 
which he appears e as subject to the laws of nature, as part 
and parcel of nature,' is not the real man, though it is 
indeed only from that point of view that man is an object of 
observation, for all that we can observe is phenomena in 
relation to phenomena. 

When it is said that a man who can only act under the 
idea of freedom, or who thinks of himself as free, is really free, 
what is meant is that a man's act is determined by what the 
man is, and what the man is is determined by this idea of 
himself as free, i.e. by the conception of himself as the object 
for which he acts. A man not merely acts so as to satisfy 
himself (probably he does not so act), but his act is deter- 
mined by the idea of himself as the object for the sake of 

1 Krit. d. r. V. p. 384 ; p. 343, Tr. 


which the act is done, and for that reason he imputes it to 
himself, and is in this sense really free. It is not a sense, 
however, in which freedom either = power to do or not to do, 
or autonomy in Kant's sense (as = a state in which he is 
determined by the conception of himself as giver of universal 
law), but it is a sense in which freedom is opposed to the 
determination of one natural event by another, or a phe- 
nomenon by a phenomenon (for this consciousness of self as 
an object is not such an event or phenomenon), and it is a 
sense in which freedom implies a certain sort of determination 
by reason. It is through reason that man conceives himself 
as the object of his actions, but the reason is imperfectly 
communicated to him so far as he has no true conception of 
what the self is which he seeks to satisfy. And this freedom, 
though it is not autonomy, and is compatible with hetero- 
nomy in a sense (not indeed of merely natural determina- 
tion, but of determination by objects incompatible with any 
law of which man can regard himself as the author), is the 
condition of autonomy. 

93. Thus Kant is right (1) in holding that a man acting 
under the idea of freedom in the sense explained is really 
free, (2) in identifying this freedom with determination by 
reason, (3) in opposing it to all natural determination ; 
he is wrong (1) in identifying freedom thus understood with 
autonomy (though it is truly the condition of autonomy), (2) 
in writing as if heteronomous actions were not free in this sense 
of freedom, (3) in reducing determination by reason to deter- 
mination by the judgment 'I ought,' (4) in speaking as if 
man, in respect of all desires not determined by this judgment, 
were a member of a merely natural world, (5) in speaking as 
if there were really two characters in a man, empirical and 
intelligible, one determined by motives in which there is no 
freedom, the other determined by reason only in a way which 
excludes determination by motives and is free. In truth 
there is only one character, and that is not empirical, in the 
sense of consisting in a relation between observable phe- 
nomena, but which on the other hand consists in suscepti- 
bility to motives, and yet at the same time, on account of the 
nature of these motives, is rational and free. 



94. If there is anything that unconditionally should be, it 
must be something that there is a reason for wishing to do 
or to become, but which is yet not desirable as a means. We 
seem to be sure that the moral end is such an object, and the 
persistency of the assurance is shown by the fact that 
hedonistic moralists are always contradicting themselves by 
trying to represent pleasure, through an equivocation between 
the desired and the desirable, as at once the unconditioned 
good (because that which alone we actually desire), and an 
object which we should desire. As Fichte says, 'As surely 
as man is man, so surely he is aware of a necessity laid upon 
him to do something, quite irrespectively of ulterior objects, 
simply that it may be done, and to abstain from doing some- 
thing else simply that it may not be done.' On the other 
hand, how are we to explain that anything should be, that 
there is a reason for desiring anything, or that it is desira&fe 
as distinct from desired — except as a means to some ulterior 
good 9 This is the great difficulty with which idealistic 
ethics have to deal. Kant first brought it clearly into view, 
and with him we will begin. . . . 

95. The essence of Kant's solution of the difficulty may 
be stated thus. There is an object which reason originates, 
which it constitutes in or out of itself. This object it pre- 
sents to the will, which he regards as a faculty of desire, 
capable of giving reality to objects in accordance with ideas 
of those objects. Here, then, is an object which does not 
originate in desire, which is not naturally or instinctively 
desired, but yet is able to excite desire in virtue of what it 
is in itself. There is reason for desiring it; it is desirable, 
and only comes to be desired (if at all) because previously 
recognised as desirable ; yet that reason does not lie in its 
relation to any ulterior object, but in the intrinsic character 


which it derives from reason as constituted by it (as given 
by reason, not presented to it). The ' should be ' represents 
the relation of this object to desire, desirable before it is 
desired, and coming to be desired because previously recog- 
nised as desirable. The ' unconditionalness ' of the c should 
be ' represents the character of this object as given a priori 
by reason and independent of empirical conditions. 1 

96. Though the above, however, is true as a summary of 
Kant's account of unconditional obligation, and is confirmed 
by the passage referred to, any one who turned from it to read 
at large in Kant himself will notice that instead of speaking 
of a practical object he speaks almost everywhere of a practical 
law; not of an object unconditionally (or universally and 
necessarily) desirable, but of a law imposed by the reason and 
demanding unconditional (universal and necessary) obedience. 
But in fact the universal practical law on which Kant insists 
is unintelligible except as implying an object unconditionally 
good to which it is relative. It has no content, it prescribes 
nothing, except what is relative to this object. The law 
which, according to Kant, regulates the good will, derives 
its authority from the conception of a good will as an 
unconditionally good object. That which the exponents of 
Kant call ' duty for duty's sake ' is rather duty for the sake 
of the attainment of that perfect will, which in imperfection 
submits to duties, but in perfection supersedes them. The 
vindication of these statements, as an interpretation of Kant, 
must be postponed. 

97. To most readers of Kant the primary difficulty is not 
whether there can be such a thing as a practical law apart 
from a practical object to which it is relative (which at first 
sight seems to be what is meant by a law which determines 
action in virtue of its mere form), but whether reason can 
properly be said to give or constitute or originate either law 
or object at all. We commonly think of reason, in regard 
both to knowledge and morality, as having no originative 
function. In regard to knowledge, we suppose its office to 
lie simply in analysing or tracing the connection between 
facts or objects presented to us by the senses, which are 
known to us (confusedly, if not clearly) without it, and exist 
altogether independently of it. So, in regard to morality, we 
consider its office to be to consider the means to various 

1 Grundlcgung, &c, pp. 275-276 ; Tr. pp. 64-66. 


objects of desire, given independently of it, and to compare 
the loss and gain incidental to the attainment of one with 
the loss and gain involved in the attainment of another. 
Reason is not supposed to affect the desirability of an object, 
or the desire actually entertained for it, except so far as it 
brings iuto view the relation in which a desired object stands 
to other desired objects to which its attainment contributes, 
or which, on the other hand, must be foregone in its attain- 
ment. Hume merely put the received English view on the 
matter in rather a more naked way than people like to have 
their views put, when he said, ( The reason is and can be 
only the slave of the passions.' With Kant, on the other 
hand, reason has a constructive or originative function. As 
applied to the sensuous receptivity, to the data of sense pro- 
perly so called, which, as Kant rightly held, apart from 
the action of reason do not amount to facts or objects 
at all, it makes the nature that we know (the 'cosmos 
of our experience'), makes the data of sense into such a 
nature. In this application it is called ' understanding,' 
which, according to Kant, it must always be remembered, 
is merely a particular exercise of reason (particularised by 
the material on which it is exercised). Again, as practical, 
i.e. as applied to the will, which Kant considers a faculty 
of giving reality to an object corresponding to an idea of 
the object, reason makes morality. It gives that conscious- 
ness of an object of which morality is the realisation. If we 
would estimate fairly this view of Kant, we must get rid of 
the special associations connected with the term ' reason.' 
Much of the current criticism of his view seems merely 
verbal. His adoption of the term ' reason ' to express what 
he takes it to express is determined by processes which the 
history of philosophy explains, but with which we need not 
trouble ourselves here. The primary question is, whether 
there is any act or function of the conscious subject, or of 
the soul as possibly acting without consciousness, or of 
anything in the nature of such conscious subject or soul, in 
virtue of which, on the one hand, the feelings received directly 
or indirectly through the bodily organs (the nervous system) 
become a connected world of experience or a knowledge ; and/ 
on the other hand, the impulses or volitions, which express! 
themselves through the bodily organs, become actions distinc- \ 
tively moral, actions to which the terms 6 ought ' or ' ought 


not,' c should be ' or ' should not be/ are applicable. If 
there is, and if such act or function or principle is not itself 
a result of feelings and impulses on which it afterwards 
reacts, then the main point in Kant's doctrine, the point 
which separates him from the English psychologists, is 
established. Whether the act or function or principle is best 
called ' reason,' is a secondary question. 

98. I do not propose here to enter with any detail into 
Kant's theory of knowledge. His method is to analyse the 
conditions of its possibility. As such a condition he discovers 
what he variously calls a function of unity, of synthesis, of 
objectivity, of necessary connection. It is this function that 
is implied in the fact that all phenomena are presented 
to us as belonging to a single order, each being qualified 
by relations which form its objective reality and neces- 
sarily connect it with other phenomena. Without the 
exercise of this function there would be no experience, no 
knowledge, and (for us) no nature at all. It cannot be the 
result of experience, since it conditions all possible experi- 
ence. It cannot be through experience that we learn to 
regard the world as one, since, unless we so regarded it, 
there would be no experience of a kind which could enable 
us to conceive a world (things, facts, or objects) at all. This 
function, or the faculty relative to this function (it makes no 
difference which we say), implied in the possibility of sucli 
experience as can yield knowledge of nature, Kant calls 
' understanding.' It is the source or bond of the relations by 
which we hold phsenomena together in one world or system ; 
such relations as those of quantity (extensive or intensive), 
of substance and its modes, of cause and effect and reciprocal 
action. But he notices also that while a process of unifica- 
tion or synthesis is implied in the consciousness of successive 
phsenomena as forming an objective world or nature, this 
unification is never complete. We never can know the world 
as a whole, never know it in its beginning or end, in its 
ultimate elements, or as a complex to which nothing can be 
added. This, according to Kant, is due to certain conditions 
of the phenomena which form the material on which the 
unifying function is exercised, viz. to their being presented 
either in space, and thus as each having another beside and 
beyond it, or in time, and thus as each having another before 
and after it. Thus the unifying principle, as the basis of 



knowledge, is for ever presenting demands which cannot be 
fulfilled, problems which cannot be solved. In respect of 
these demands or problems as insoluble Kant calls it reason 
in a special sense, as distinct from the understanding, which 
is a name for the same principle in so far as its work appears 
in that partial fulfilment of its demands which constitutes 
our (ever and necessarily incomplete) knowledge of nature. 
Those who imagine that reason can tell us one thing, under- 
standing the contradictory, may be right or wrong, but they 
do not represent Kant. The judgments or conceptions, which 
he considers the work of understanding, are just as much a 
result or expression of reason as the quest after the absolute 
and unconditioned ; but they are its expression as striving 
after completeness of synthesis where that completeness is not 
to be found. The demand from which they result is not satisfied 
by them, but on the other hand not contradicted by them ; 
for in the nature of the case a conception or judgment can 
only contradict another conception or judgment, not a 
demand. The judgment, e.g., that every event is determined 
by a preceding event, arises from the action of reason as a 
principle which compels us to seek to hold phenomena 
together as elements in an interrelated whole (to seek for 
complete synthesis or unification of phenomena) ; while at 
the same time this necessary way of regarding phenomena 
as each an event determined by a preceding event, is incom- 
patible with our conceiving them as a whole or knowing 
their whence and whither. There does not follow from this, 
however, any conflict between reason and understanding. 
It only seems to follow if we unwarrantably convert the 
demand or effort of reason to find a unity, and so to regard 
every phenomenon as related to another, into the proposi- 
tion that phenomena form a limited series, and that there 
has been an absolutely first event (a proposition which 
conflicts with the judgment of understanding stated above) ; 
or if, on the other hand, we construe this judgment, in fact 
only applicable to phenomena and arising out of the action 
of reason upon them, as if it applied to reason itself, and on 
the strength of it treat the principle, in virtue of which alone 
we regard all phenomena as conditioning each other, as if it 
were itself a conditioned phenomenon, and its demands 
illusions, to be explained like any other illusions, by inquiring 
into their natural history. 

99. Readers of Kant often find a difficulty in tracing any 


real identity between the ' reason ' of which he speaks 
in the Critique of Pure Reason and that of which he 
speaks in his moral treatises. In the Critique of Pure 
Reason they find reason treated as the source of certain 
' ideas,' ideas of an * unconditioned ' under three forms, 
severally implied in the search for an ultimate substance, 
for completeness (a beginning and end) in the series of 
events, for a unity in which all realities are included ; ideas 
which regulate all knowledge, but to which no knowable 
object corresponds. What can such ideas have to do with 
moral action ? What has the reason, which is the faculty 
or source of them, in common with the consciousness of an 
object or law of conduct, or with the autonomous will 
which Kant in his ethical writings identifies with reason? 
By ideas we 'understand conceptions of existing objects or 
phenomena under their relations or general attributes, and 
this gives us a false notion of what Kant means by the 
\ ideas ' of which reason is the faculty. His meaning is 
better expressed when he calls reason the source of problems, 
of operations which it sets us upon performing. The ideas 
which he ascribes to reason are not ideas derived from 
objects, or to which any object can be found to correspond : 
they are ideas which strive to realise themselves, and in so 
doing result in the functions of the understanding. Thus 
the idea of a totality of conditions, given by reason, expresses 
itself in the process of connecting each phenomenon with 
another as its condition, though through that process such 
totality can never be reached. It is a demand, arising out 
of the nature of reason, that such a totality be found. Thus 
the ideas of reason, even in its theoretical application or in 
its relation to knowledge, are ideas of what should be, which 
are in a way realised in the direction given to the processes 
of knowledge, but not in a knowable object ; (not in a know- 
able object, because a sensible matter is necessary to know- 
ledge, and the ' forms of sensibility ' prevent correspondence 
with, or realisation of, the idea or demand of reason). And 
just as reason, through the idea of a totality of conditions 
which it supplies, gives direction to the process of know- 
ledge, so through an idea of like origin, which Kant com- 
monly regards as the idea of freedom, i.e. l of a subject 

1 Krit. d. r. V. p. 266 ; Tr., p. 229. He speaks of it as ■ the idea of the 
necessary unity of all possible aims.' 


determined solely by reference to itself as an absolute end, 
it gives direction to the practical powers. 

100. There is, however, this essential difference between 
the two applications of reason. When the idea of reason 
is applied theoretically (in the direction of knowledge), the 
question whether it is realisable in any object of the senses, 
whether any such object can be found answering to it, is 
appropriate, shuse knowledge is necessarily concerned with 
such objects, as a process of bringing them into relation, 
and is otherwise a mere play of words ; and it is a question 
which, as we have seen, must be answered in the negative. 
On the other hand, in regard to the practical application of 
reason or its ideas, such a question is unmeaning. The only 
question here is, whether it is realisable in a determina- 
tion of the will, a question akin to the question whether it 
directs the process of knowing or effort after knowledge, as 
distinct from the question, whether any knowable object 
can be found corresponding to it. Thus the expression, ' it 
is only an idea ' (as implying that no phenomenon corre- 
sponds to the idea), very significant when a question of 
knowledge or matter of fact (which implies relation to the 
senses) is at issue, has no force as depreciatory of a principle 
of action. If such an idea or principle can affect the will, 
that is all that from the moral point of view is to be ex- 
pected of it. The moral relation is of idea to desire, not 
(like the cognitive relation) of conception to sensible object. 
If a certain idea (Kant would say the idea of freedom, or of 
oneself as imposing law on oneself) so affects a man's 
desires that his strongest desire (or will) corresponds to it, 
all moral requirements are satisfied. The fact that the 
sensible objects resulting, the effects of the will in the 
world of phenomena, are not, from their nature as sensible 
objects, cognisable as corresponding to the speculative idea 
of freedom (the idea of a causa sui), any more than any 
sensible objects can be found to correspond to that idea of 
reason which governs the operations of the understanding 
in any of its forms, has no bearing on the question whether 
or no the idea of freedom, as practical, determines the will. 

101. Thus the Critique of Pure Reason prepares the 
way for Kant's ethical doctrine (1) in exhibiting the action 
of what he would call an * a priori law-giving ' faculty as 
the condition of our consciousness of nature, of a world of 


objects, through analysis of experience, and detachment of 
the pure element in it; i.e. the element which cannot result 
from experience because necessary to there being an experi- 
ence as of objects. The sense in which this faculty is called 
' law-giving ' is indeed quite different from that in which 
we speak of moral laws being given. It ' gives laws ' in the 
sense (a) that it determines the condition under which we 
can know a nature (under which phenomena become to 
us a nature) ; (b) that from it results the consciousness of 
there being laws of nature. But when we have guarded 
against any misuse of the term c law-giving,' it remains true 
that there is in man, over and above his sensuous recep- 
tivity, a power of objectifying his feelings as related facts, 
which renders them into a knowledge. Thus his theoretic 
activity, the fact that there is such a thing as knowledge, 
a consciousness of relations, evinces the action upon sensuous 
data of a thinking principle not derived from them. (2) It 
appears further from the Critique of Pure Reason that this 
thinking principle, in its' theoretic activity or as exercised in 
knowledge, not only cannot satisfy itself, but that the judg- 
ments, universal and necessary, as to the relations of 
phenomena which result from its application to phenomena 
(under the name ' understanding ') are incompatible with 
its ever finding in them an object answering to the idea of a 
possible completeness which it brings with it. When Kant 
has established, then, at once the existence of this transcen- 
dental principle (i.e. a principle not a result of sensuous 
experience) as the condition of the possibility of knowledge, 
and the limitation of its exercise in knowledge (which with 
Kant is only possible in relation to phenomena), the way is 
cleared for the inquiry whether there is any other exercise of 
it in which it is not so limited ; an exercise in which, instead 
of having to do with a material determined (a) by an unknown 
cause beyond ourselves, (b) by the necessary forms of our 
perceptions, space and time, it is applied to a faculty of 
bringing objects into existence corresponding to ideas. 

102. This inquiry Kant describes as the inquiry whether 
there 'is a practical use of pure reason.' Practice with 
him consists in giving reality to conceived objects, or bring- 
ing objects into existence corresponding to ideas. That 
there is such a thing as practice in this sense, however it 
may be explained, does not need to be proved. Having 


thought of an object consisting (say) of a service to a friend, 
or a hurt to an enemy, or of a gratification of some taste of 
his own, a man proceeds by act of will to give reality to the 
object thought of. The act of will (' Handlung ') may not 
be successful for its purpose ; the man, having tried to serve 
his friend, &c, may fail to bring about the combination of 
natural events or circumstances without which the benefit 
does not accrue to his friend ; still he has not merely 
thought but acted, not merely wished but willed, not merely 
desired but intended, and in Kant's sense has given reality 
to his idea, has brought an object into existence corresponding 
to it, though there is an ulterior sense in which his idea 
remains unrealised. This distinction between the realisation 
of an idea in act or determination of will, and its realisation 
in some ulterior outward effect of the act of will, is important 
for the understanding of Kant. 

There being no doubt, then, of the fact of practice as a 
realisation of ideas (' Yorstellungen'), in the wide English 
sense of ' ideas,* the question remains as to the origin of the 
ideas thus realised. Do they in any case originate in reason, 
as the transcendental principle of which the action was 
found to be implied in knowledge, but to the demands of 
which no known object answers ? Are they ever the product 
of pure reason, or has reason as pure anything to do with 
their production? We say pure reason, because no one 
doubts that in some sense or other reason determines the 
will, that we act or may act on ' rational grounds,' do so 
and so because our reason tells us. The question is whether, 
in so determining will, reason is ' pure ' or ' empirically 
conditioned,' i.e. conditioned by antecedent desire, with the 
constitution of which it has nothing to do ; in other words, 
whether it can be an ultimate and unconditioned determinant 
of action. (Cf. the opening of the Introduction to the Critique 
of Practical Reason.) Thus Kant constantly puts as the equi- 
valent to the question, ' Is there a practical use of pure reason ? ' 
the question, ' Does reason exercise a free causality, i.e. an 
absolutely originative causality ? ' which again he often 
puts simply as the question, ' Is there a free causality ? ' 
For it is self-evident that the attribute of freedom, of 
being an unconditioned source of conditions, if it is to be 
found in man at all, can alone belong to that principle in 
him in virtue of which he regards everything as conditioned. 


If anything is not natural (the natural and conditioned being 
equivalent), it must be that through which we conceive 
nature (or, as Kant sometimes puts it, make nature, in the 
sense that through it alone there is for us a nature). 1 

103. The answer to this question in both its forms, 
according to Kant's view, is furnished by a fair interpretation 
of what we call the ' voice of conscience,' our actual moral 
judgments (' gemeine sittliche Vernunfterkenntniss '). The 
idea of duty cannot be the result of a generalisation from 
sensible experiences, of desires for pleasure and observation 
of the means to gratify them. It must originate in reason. 
If, then, there is such a thing as determination by this idea, 
there is a free causality, a practical use of reason as pure. 
As he puts it, though freedom of will is the ratio essendi to 
moral obligation (since it is the freedom of the will, i.e. its 
relation to pure reason as determinable by it, that makes 
such a thing as action from the idea of duty possible), it is 
the consciousness of obligation and the effect of such con- 
sciousness on action that is the ratio cognoscendi to the exist- 
ence of freedom. It is difficult indeed to see what difference 
there is in this case, according to his doctrine, between the 
ratio essendi and the ratio cognoscendi. Freedom of will, as 
he understands it, is not something different from a capability 
on the part of the will of being determined by the idea of 
duty (or by reason, which is the source of this idea), and 
proved by such capability. It is the capability itself. One 
and the same spiritual condition, described negatively as 
freedom, i.e. described as absence of determination by merely 
natural causes, is described positively as determination by 
the idea of duty. As Kant puts it, ' die Unabhangigkeit von 
aller Materie des Gesetzes (namlich einem begehrten Objekte) 
. . . ist Freiheit im negativen, diese eigene Gesetzgebung 
der reinen und, als solche, praktischen Yernunft, ist Freiheit 
im positiven Verstande.' 2 Kant, however, scarcely seems 

1 In the Introduction to the Critiqtie in the will but that which belongs to it 
of Practical Bcason Kant writes as if in virtue of its determination by reason, 
the questions (a) Has reason as pure a and by reason as pure, i.e. as anultimate 
practical use? and (b) Does freedom determinant, an absolute first, 
belong to the will ? were different 2 ' The independence on all matter of 
questions, and as if an affirmative an- the law (namely, a desired object), . . . 
swer to the first were to be derived is freedom in the negative sense, and this 
from the affirmative answer previously self-legislation of the pure, and there- 
given to the second. But, in fact, as fore practical, reason is freedom in the 
he treats them they are equivalent positive sense.' — Krit. d. p. V. p. 35 ; 
questions, for he recognises no freedom Tr., p. 170. 


fully to realise Lis own identification of freedom with deter- 
mination by reason. Here as in other cases he begins with 
a recognised distinction which his own doctrine explains 
away, but which he continues to treat as if it were valid. 
At any rate his method is to begin with establishing the 
reality of determination by the mere idea of duty, and then 
to deduce freedom of will from this. 1 The first step is taken 
in the Grundlegung zur Metaphysik der Bitten of which, as 
Kant says in the preface to the Critique of Practical Reason, 
the object was to clear up the principle of duty and establish 
a definite formula for its expression. It is difficult at first to 
understand the difference between the questions which these 
two treatises are severally meant to answer. The problem 
of the latter, according to Kant's statement, is to show that 
and how pure reason can determine the will, and give reality 
to objects so far as is implied in such determination. It is 
thus parallel to the Critique of speculative reason, of which 
it is the problem to exhibit the function of reason in render- 
ing knowledge possible. The Grundlegung, on the contrary, 
does not ostensibly deal with any question in regard to the 
function of c pure reason.' It examines the idea of a good 
will, as that which the unsophisticated man recognises as 
the only thing of absolute value in the world. It finds that 
this good will is will determined by the idea of duty ; not 
merely a will which conforms to duty, but to which duty is 
the motive. This means a will determined by the concep- 
tion of a law which commands without possibility of ex- 
ception, and not hypothetical^ bub categorically; not in the 
form, ' act thus under such and such circumstances if you 
wish to get such or such a pleasure,' but \ act thus under all 
circumstances and for the sake of acting thus.' Such a 
law, categorical and universal, can only be one imposed by 
the agent on himself, in virtue of a nature which renders him 
an absolute end and expresses itself in a law equally binding 
on all agents capable of a consciousness of law. Thus the 
result of the Grundlegung, starting from the idea of ' good 
will' as established in the common conscience, is to exhibit 

1 (1) Is there determination by the question in the first form, and not in 

mere idea of duty ? (2) Is there a free the two latter. Thus, though the an- 

causality, belonging to will ? (3) Is swer to it in the first form really in- 

there a practical use of pure reason ? volves the answer to it in both the 

These three are, in fact, according to others, Kant sometimes writes as if the 

Kant, equivalent questions, but we have latter answers were inferred from the 

the means of directly answering the former. 


this as the autonomous will, i.e. as a will of which the rule 
is ' given by the rational subject himself in his character 
as giver (originator) of law universal.* In the Critique of 
Practical Reason Kant does nothing further to justify or 
establish this view of what is involved in pure morality, 
though he somewhat varies the form in which he repeats it, 
but he applies it to the solution of the problem stated above 
(to show that and how pure reason can determine the will, 
&c). If man can give to himself an absolute and uni- 
versal law, so that his conduct is affected by it, pure reason 
can determine the will, for only pure reason can yield such a 
law, a law not dependent on any kind of sensuous experience. 

[Here follows a sketch of the main points arrived at 
in the Grundlegung and part of the Critique of Practical 
Reason. The results of the Grundlegung are thus summed 
up : '■ Good will is will (1) determined by a categorical im- 
perative, i.e. an imperative not conditional upon the desire for 
any end ; (2) will according to which a man makes humanity 
in his own person and that of others his end ; (3) autonomous 
will, i.e. will so acting that it can regard itself in its maxim 
as giving universal law.' The lecture then proceeds to ex- 
amine Kant's doctrine (Grundlegung, 2ter Abschnitt), that 
the idea of duty is no ' Erfahrungsbegriff.'] 

104. Kant's point is that an idea of what should happen 
cannot be derived from, or consist in, a generalisation as to 
what does happen ; therefore is not an ' Erfahrungsbegriff.' 
This, so far as it goes, is scarcely open to dispute. The 
proposition ' so and so should or ought to be done,' is not 
a proposition as to a matter of fact at all, and cannot be 
reached by the same process by which we arrive at general 
propositions concerning matters of fact. (Hume noticed 
this, Treatise of Human Nature, bk. iii. pt. i. § 1.) But 
though not a result of generalised observation, does it follow 
that it is not a product of experience, of an experience which 
not merely informs us as to matters of fact, but determines 
the direction of our desires and aversions ? Such experience, 
it must be conceded, cannot yield a ' categorical imperative,' 
an imperative independent of every desire and fear on the 
part of the subject conscious of it : but we need not assume 
the consciousness of such an imperative in order to explain 
judgments expressed b}- 'ought' and 'ought not.' Simple 
natural desire or aversion, indeed, cannot account for the 


consciousness in question. To say that I ought to seek 
pleasure, as a moment's reflection shows, is absurd. But 
experience (a) causes us to associate certain sequent pains 
with the attainment of certain pleasures, (b) produces a 
conflict of desires by making us aware that to gratify one is 
to forego the gratification of another, (c) leads to certain 
requirements on the part of society, or those who wield the 
forces of society, founded on observation of what is for their 
interest, which are embodied in law or convention. Hence 
arises the consciousness represented by the judgment should 
be or ought to be, which is reducible to (a) ■ I had rather not, 
for by so doing I lose such a pleasure, but I must, in order 
to avoid such a pain, or to gain such another pleasure,' or 
(b) ' I had rather not, but I shall be punished if I do not,' 
which, through education, personal or hereditary, becomes 
the judgment (c) ' I had rather not, but it is expected of 
me, and the pain of not doing what is expected of me out- 
weighs the pleasure which I lose by the action,' from which 
any distinct apprehension of punishment to be avoided has 
disappeared. It must be noticed, however, that the form of 
consciousness thus accounted for is merely a desire for some 
pleasure or aversion from some pain, modified in a particular 
way by conflict with desire for some other pleasure or aver- 
sion from some other pain. It is not a consciousness on the 
part of the individual of an obligation irrespective of his 
inclinations. A consciousness of obligation which is so 
irrespective, though it be an obligation to pursue one's 
own greatest happiness, cannot be thus accounted for. The 
question at issue relates, not primarily to anything that 
we are obliged to do, but to the nature of the obligation 
to do anything. If there is an obligation, and a con- 
sciousness of obligation, to pursue my own greatest happi- 
ness, whether I desire it or no (whether it is pleasant 
to me or no), it must be explained in some other way 
than by any modification, through experience, of desire 
for pleasure or aversion from pain. And we have already 
seen that experience, as relating to matters of fact, can have 
no bearing here. The judgments which it produces or 
modifies are essentially different from the consciousness in 
question. Any notion to the contrary arises from a con- 
fusion between the consciousness of obligation and the con- 
ception of the object to which the obligation relates. Given 


a consciousness of obligation to one's neighbour, experience 
as to matter of fact may affect the judgment as to who one's 
neighbour is. Nay, if we push the inquiry further back, we 
may find that without such experience there would be no 
conception of a neighbour at all as an object in relation 
to which any obligation exists. But this does not affect the 
question as to the origin of the peculiar form of conscious- 
ness called consciousness of obligation. In like manner the 
observations which lead some people to ascribe this con- 
sciousness to the progressive experience of utility, are beside 
the point. Through growing enlightenment, derived from 
experience, men may come to recognise a duty to do that 
which they did not recognise as a duty before, and to abstain 
from that which they thought it a duty to do; but the 
recognition of duty, the capacity for judging, c I ought to,' 
and being determined to act by the judgment, remains 
unaccounted for. It is, indeed, more logical to deny the 
existence of the consciousness of duty, in the proper sense, 
as a factor in the determination of conduct, and treat it as a 
disguised fear or as a result of conflicting desires and fears, 
than, admitting its existence, to treat it as the product of 
the experience of utility. The real question, then, is, in 
Kant's technical language, whether there is such a thing as 
a ' categorical imperative ' at all, or whether there are only 
* hypothetical ' imperatives ; in other words, whether there is 
an element in the formation of character and determination 
of conduct consisting in a consciousness of the desirable as 
distinct from the desired (a consciousness of an object which 
determines desire instead of being a result of the desire 
for pleasure), or whether, on the other hand, the conscious- 
ness of obligation being ultimately dependent on desire for 
pleasure, each obligation is conditional upon a preponderance 
of pleasure accruing in the result, and thus upon the sus- 
ceptibilities to pleasure and pain of the individual obliged. 
This is not the place to discuss this question fully. Perhaps 
it is sufficiently answered by being stated clearly. At any 
rate those who profess to adopt the latter answer seem 
always somewhere to avail themselves of a distinction be- 
tween the desirable and the desired (to assume a practical 
consciousness of the desirable as distinct from actual desire 
for pleasure), to which they are not entitled. What we 
have here to point out is that Kant, adopting the former 

124 KANT: MORAL nilLOSOniY. 

answer, is quite right in refusing to regard the c categorical 
imperative' (or sense of duty as he understands it), the 
consciousness of the unconditionally desirable, as a product 
of experience, and that the ordinary judgments as to what 
is desirable are not to the purpose. 

105. It is another question whether the a priori factor in 
our moral consciousness, which can alone give any moral 
bearing to the experience of matters of fact and that of 
pleasures and pains, is properly an idea of universal law at 
all. The conviction that it only gradually takes this form 
strengthens the repugnance to Kant's doctrine, that the 
true principle of morality is independent of experience. He 
is quite right, indeed, in insisting that the consciousness of 
a universal practical law is not a product of experience in 
either of the above senses ; but this should have led him to 
find its origin in a principle of consciousness which makes 
experience what it is, rather than to treat it as an absolutely 
original datum. As it is, though his doctrine is essentially 
true, his way of putting it excites the same opposition as 
his way of putting the corresponding doctrine in regard to 
the a priori element in knowledge. He is quite right in 
insisting that what he calls the a priori principles of science 
(e.g. ' all changes take place according to the law of the 
connection of cause and effect ') are not of empirical origin, 
in the sense that they are not a possible product of a 
succession of sensations. They are only evolved from 
experience because they are already involved in it (in that 
sort of experience which is the parent of knowledge). They 
are not a product of experience, but are produced in and 
with experience by that action of a self-conscious subject 
upon the succession of feelings which constitutes experience. 
It is only, however, through reflection upon experience, 
which can scarcely supervene upon it till it has reached an 
advanced stage, that they can be recognised in their abstract 
character as its universal and necessary laws ; whereas Kant 
is apt to be understood (partly by his own fault) as mean- 
ing that they are an c a priori furniture ' which the mind 
possesses antecedently to experience. In like manner the 
consciousness of an absolute moral law is not a product of 
desires for pleasure and aversions from pain, however 
modified by the natural association of ideas ; nor, if our 
moral experience were a succession of such desires so modified, 


could it ever arise in it. But in fact our moral experience 
is what it is only through that action of the self-conscious 
subject upon desires, from which arises the practical idea 
in various forms of an absolutely desirable object. Moral 
experience thus constituted, there arises through reflection 
on it in a certain stage of development the practical idea 
of an absolute moral law (' practical ' meaning throughout 
6 determinant of action '). This idea may then rightly be 
called a priori in the sense that it only arises in moral 
experience because that experience implies the presence of 
a ' non-sensuous ' agent, a self-consciousness not reducible 
to, or produced by, any number or kind of desires, but 
which renders desires into elements in a moral character. 
It is a mistake to call it so in the sense that in this abstract 
and recognised form it is given in the moral experience of 
man to begin with; still more so in the sense that it is 
separable from that experience and antecedent to it ; nor is 
this what Kant really meant to convey. The term ' a priori 9 
with him conveys no reference to time, implies no ante- 
cedence in time on the part of that which is called ' a priori ' 
to that which is opposed to it. The conceptions which he 
calls ' a priori * in his theory of knowledge are the a priori 
element in experience (not prior to it), which again = the 
6 pure ' or non-sensuous element, the element not derivable 
from any succession of feelings. It is in a corresponding 
sense that the idea of absolute law is an a priori or pure 
element in moral experience, or, in Kant's language, a pure 
practical cognition ('praktische Erkenntniss '), i.e. the con- 
sciousness of an object not derived from the senses ( = pure) 
by which the will is determined to make that object real 
(= practical). If he is asked what the object of the con- 
sciousness described is, he answers that it consists in a will 
determined by the consciousness (idea) of universal law as a 
law imposed by itself. Consciousness of a possible object, 
consisting in a will determined by the idea of universal 
law, tends to make that object real, to determine the will 
by such an idea. 

106. If Kant is asked further what is the origin of this 
consciousness, to which a sensuous origin is denied, he 
answers ' reason ' ; but of this he tells us no more than that 
it originates an activity directed to the realisation of itself, 
or to the attainment of a rational condition, i.e. a condition 


in which will, as the faculty or possibility of being deter- 
mined by the consciousness of law, is actually so deter- 
mined ; or, again, that it is the source of the practical idea 
of freedom, i.e. of an idea of self-determination which tends 
to realise itself in the subjection of man to a law which 
he imposes on himself. It is clear, however, that we learn 
nothing more about the moral consciousness, when it has 
been analysed as Kant analyses it, by referring it to a 
faculty called reason, any more than we learn anything 
about any function by referring it to a faculty, unless we 
know something of the faculty which is not contained in 
the analysis of the function. If the analysis of our moral 
experience shows that character is actually affected by the 
moral agent's consciousness of himself as capable of being 
determined by the conception of a self-imposed law as such 
(as apart from any object to which this law prescribes the 
means), we may state this in another form by saying that 
there is a faculty called reason, which originates an activity 
directed to the realisation of itself; but this is merely to 
give a name to the consciousness already described, unless 
it enables us to connect the function thus ascribed to reason 
with other functions of it, e.g. with its functions in render- 
ing possible science and art ; and in his moral treatises 
Kant scarcely attempts so to connect it. So far 3 therefore, 
as the Critique of Practical Reason merely adds to the 
analysis of the moral consciousness, given in the Grund- 
legung, a derivation from reason of the e imperative ' which 
the moral consciousness has been found to involve, it does not 
seem to contribute anything to the theory. The important 
question is the truth of the analysis. Is the distinctive 
thing in the moral consciousness rightly held to be the 
presentation to the moral agent of an end or object consist- 
ing in his self, as determined by a self-imposed law, in virtue 
of its mere form as law and self-imposed ? 

107. Such a presentation is the basis of the c categorical 
imperative ' in each of the three forms in which Kant states 
it. (1) ' Act only according to that maxim by (' durch,' in 
another place, ' von ') which you can at the same time wish 
that it (the maxim) become a universal law ; ' (otherwise 
put thus, ' Act as if the maxim of your act were destined 
through your will to become a universal law of nature ; '). 
(2) ' Act so as to treat humanity, whether in your own 


person or in that of anyone else, always as an end ; if as a 
means, still as at the same time an end (this is the force of 
the ' zugleich'), never merely as a means.' (3) ' Act only in 
such a way as that the will can regard itself by its maxim 
as at the same time giving universal law.' 

The maxim is the recognised motive of an act (a 
strongest motive which the agent recognises as such). If a 
man's strongest motive at any time is to get money, so that 
he says to himself, ' I must get some money,' then that is his 
' maxim.' If my motive is the wish to obey a law simply 
because it is a universal law and one which I impose on 
myself, or to bring myself into a state in which I shall habi- 
tually do so, then (a) I am acting from a motive which I can 
at the same time wish to be universally acted from, by 
myself at all times, by all beings capable of conceiving such a 
law at all times. On the other hand, if the motive is any 
other kind of wish, a wish determined by anything but the 
conception of oneself as giving a law to oneself, it is accord- 
ing to Kant a wish for some pleasure to oneself, and a wish 
for pleasure to oneself cannot at the same time be wished to 
be universal law (be wished to be the wish or motive of 
everyone else). While wishing pleasure to myself, I may 
hold the opinion that all men wish pleasure for themselves, 
but that is quite a different thing from wishing my wish to 
be wished by everyone else. One may present it to oneself 
as a fact that everyone wishes for pleasure; one cannot 
present it to oneself as desirable, any more than any other 
physical fact, say, that everyone digests food. Further, if 
my motive in any act is as above, then (6) I am ' treating 
humanity in my own person as an absolute end,' for my 
ultimate object is to fulfil myself in respect of that faculty 
or possibility which is distinctive of me as a man. It is the 
conception of myself (as Kant would say) ' in the dignity of 
a rational being obeying no law but that which he himself 
also gives,' that determines the act. On the other hand, if 
my motive in any act is desire for pleasure, 1 (and according 
to Kant's view — of which more afterwards — there is • no 
alternative between the motive consisting in the wish to 
obey a self-imposed law and that consisting in a desire 
for pleasure), I am treating 'humanity in my person' as 
a means to an end which the mere animal susceptibility 

1 Cf. J. S. Mill, Utilitarianism, p. 13. 


constitutes. Again, in any act motived by the wish to con- 
form to a self-imposed law, or to bring myself into a state 
in which I shall habitually do so, I cannot be 'treating 
humanity in the person of others ' merely as a means, for 
my object is one common to myself with all others so far as 
properly human (or, as Kant says, so far as they have a 
rational nature), and thus capable of the same motive. It 
must not be objected that, if according to a natural law all 
desire pleasure, pleasure is equally a common object. To 
desire pleasure is to desire one's own pleasure. To desire 
someone else's pleasure, unless it be as a means to one's 
own, is not to desire pleasure at all. Therefore, in desiring 
pleasure, every man is desiring an object private to himself, 
and so far as he uses anyone else in obtaining that pleasure, 
he is using ' humanity in the person of the other ' merely as 
a means. On the other hand, so far as I seek to conform 
myself to a self-imposed law, just for the sake of conforming 
to it, there is nothing private to myself in the object. The 
form of law, as distinct from any particular matter to which 
the law relates, is the same for all rational beings, reason 
being just the faculty of conceiving law; and in thinking of 
myself as a giver of law to myself, I necessarily think of 
every other rational being as doing the same, and thus of 
myself as a member of one community with them. If, then, 
it is really the idea of conformity to self-imposed law which 
moves me to any act, I cannot in the act be using humanity 
in the persons of others otherwise than as an end. The 
conceived end to which I seek to give reality, is equally an 
end to everyone else. It is in my own person that I seek 
to realise it, but in so doing I am realising it for the benefit 
of everyone else, for everyone is concerned in the disinte- 
rested readiness to conform to law on the part of everyone 
else. Such conformity on the part of everyone else I must 
desire in desiring it for myself, and everyone else in desiring 
it for himself must desire it for me. It is an object to us 
not in virtue of that element in our nature (the desire for 
pleasure) which separates us, but in virtue of that which 
unites us, which gives us the idea of a common good, our 
reason ; for my conformity to such law is equally an object 
of interest to them so far as they are rational, and no less so 
is their conformity an object of interest to me. 

108. That an act done from the motive described corre- 


sponds to the categorical imperative in the third form in which 
Kant states it, hardly needs to be pointed out. This third 
form of putting the imperative is directly adapted to the 
' principle of autonomy,' the principle that the moral will is 
the will which is a source of law to itself. If the motive to 
any act is desire for pleasure, I cannot regard myself in the 
maxim of the act (which will be ' seek pleasure, or so and so 
as a means to pleasure ') as giving, or as a source of, uni- 
versal law. In such an act I am taking my rule of conduct 
from a natural inclination, and in so doing I cannot regard 
myself as the author of the rule of conduct to which I con- 
form, any more than I can regard myself as the author of 
any natural process. I may give the form of self-imposed 
law to a method of action founded upon the desire for plea- 
sure. Having found that the most pleasure is attained by a 
certain course of action, I may make it my rule to follow 
that course. Still it is the ' matter ? of such a rule, i.e. the 
pleasure which it is calculated to obtain, not its / form ' as a 
general rule, on which the e maxim ' of the corresponding 
act depends, and therefore I cannot regard myself, in setting 
such a maxim before myself, as giving, or as author of, law. 
Nor is a ' law,' which represents the means by which I pro- 
pose to myself to attain certain pleasures, a universal law at 
all. It may be that all others would best obtain pleasure by 
the same means. Still as the end, my own pleasure, for the 
sake of which I adopt such a rule of action, is private to 
myself ; I can only set the rule before myself as a rule for 
myself alone. No one can be conscious of a rule, only 
binding on him as tending to his pleasure, as in its own 
nature binding on others, though he may have reason to 
think that each of those others presents a corresponding 
rule to himself as tending to his own pleasure. The only 
object in willing which I can regard my will as the source of 
law is my own conformity to law. It is the only object of 
which it can be truly said that nothing natural in me, but 
merely that which constitutes me a man, a person, in dis- 
tinction from a part of nature, that which enables me to say 
' 1/ without contribution from any other element, renders it 
possible for me. So far, then, as this object supplies the 
motive to any act, so far as my strongest desire is desire to 
obey a law without any ulterior object, I can in recognising 
the motive (in presenting it to myself as my maxim) at the 



same time regard myself as in it an ultimate source of law, 
and that a law which, for the same reason that I regard 
myself as the author of it (the reason, namely, that it arises 
out of the pure principle of personality), I must regard as a 
law for all other persons. 

109. It is thus his conception of the good will as the 
autonomous will that is the basis of Kant's Ethics. About 
this the following questions arise, (1) Is there such a thing 
as a determination of the will by the mere idea of conformity 
to law, or by a law of action in virtue of its ' mere form ' 
irrespectively of any object to which it prescribes the means 9 
(2) Granting that the will may be so determined, what 
truth is there in the notion that in being so determined it is 
' autonomous,' as determined by a law of which it is itself 
the source, whereas in being determined by desire for any 
other object than the fulfilment of law for its own sake it is 
( heteronomous ' ? (3) What truth is there in the identifica- 
tion of virtue with autonomy of the will thus understood, 
of vice with heteronomy ? in the view that states of the will 
can be divided into that in which it has no other object than 
conformity to a universal and self-imposed law, as such, on the 
one side, in which state it is good, and that in which it is 
determined by desire for pleasure on the other, in which 
state it is heteronomous and bad ? 

110. The difficulty which at once suggests itself is, that 
a law without an object is nothing at all ; yet it would seem 
that only by such a law, according to Kant, may the will be 
determined if it is to be good ; or at any rate that, if the 
law obeyed has any object, this object must have no influence 
in determining the good will to obey it. What is it that is 
enjoined by that law which I am to obey for the mere sake 
of obeying it ? Nothing, according to Kant, beyond merely 
obeying it but that I should have a will to obey it ; but what 
conception can we form of the will to obey it, how can we know 
whether we have such a will or no, or whether any action 
represents such a will or no, unless the law has some content, 
unless it enjoins something besides willing obedience to 
itself, by reference to which that willing obedience may be 
tested ? Exclude from the law, as Kant requires to be done, 
all relation to a e matter ' or ' an object of which the reality 
is desired,' 1 and what is left of it but a word ? Does not 

1 Krit. d.p. V. p. 21 ; Tr., p. 149. 


the notion of * duty for duty's sake/ in short, when logically 
worked out, prove self-contradictory, since it reduces itself to a 
duty to do nothing ? And is not Kant's real merit the nega- 
tive one of having worked out this notion more logically than 
anyone else, and so made this self-destructive result apparent ? 
111. We answer; when Kant excludes all reference to an 
object, of which the reality is desired, from the law of which 
the mere idea determines the good will, he means all reference 
to an object other than that of which the presentation ipso 
facto constitutes the moral law. That in that law, the 
willing obedience to which characterises a good will, there is 
implied some relation to an object, and that this object 
moves the will in the right sort of obedience to the law, 
appears from his account of man as an absolute end, on 
which he founds the second statement of the categorical im- 
perative. But it is one thing first to desire an object, of which 
the presentation does not in itself carry with it any idea of 
obligation (of a claim independent of any inclination we may 
happen to have), but, on the contrary, is itself simply a 
conscious inclination, and then, upon reflection as to the 
best means upon the whole of obtaining the object, to im- 
pose a law upon oneself to adopt those means ; it is another 
thing to be conscious of an object as desirable in such a way 
as that the consciousness carries with it the idea of a law, a 
claim on me to make the object mine whether I am inclined 
to do so or no. In the former case an antecedent desire for 
the object is the source of the rule which I come to impose 
on myself as the condition of my gratifying the desire. In 
the latter case, the consciousness of the object as having a 
claim on me, or as a source of law, is the condition of any 
desire I come to have for it, and through that of its de- 
termining my will. Kant's point is that the object which 
may alone form a ' Bewegungsgrund ' of the will, if the will 
is to be good, must be of the latter sort ; that the desire for 
it must be derived from the prior conception of it as desir- 
able. This (which he expresses by calling it a ' formal ' as 
distinct from a 'material' principle) is the condition of its 
being a source of law having ' objective necessity,' in dis- 
tinction from the subjective necessity of the individual's 
inclination. Since such an object, in its relation to us, con- 
stitutes an absolute law ; since in distinction alike from that 
for which we have a natural desire (to which the should be 


has no application), and from the means to satisfy such 
desire (which only conditionally should be), it is presented as 
that which unconditionally should be; Kant considers the 
determination of the will by it to be a determination by a 
law in respect of its c mere form ' as apart from any ulterior 
object to which the law prescribes means. The ultimate 
question, then, will be, Is there such an object, and is Kant 
right in his account of it? A further question will be, 
whether the conception of law is properly employed to 
express our relation to it, and its to us. 

112. According to Kant, 'man in his rational nature is 
an absolute end.' This at least seems to be the result of his 
two propositions ; l (1) ' man, and every rational being as 
such, exists as an end in itself;' (2) 'the rational nature 
exists as an end in itself.' His view seems to be that since 
every man, in virtue of his rational nature, presents his own 
existence to himself as an absolute end, there is for every 
man (1) a subjective end consisting in his own existence as 
an end in itself; (2) an objective end arising out of the fact 
that everyone else presents his own existence to himself as 
an absolute end, and consisting in the rational nature as 
common to each man with everyone else capable of present- 
ing his own existence to himself ; and (3) that to the good 
man or good will, this ' objective ' end becomes identical 
with the ' subjective ; ' so that it is only in respect of a rational 
nature or humanity, common to himself with all others, that 
he is an absolute object to himself. 

But hereupon it may be asked, (1) With what right 
is it assumed that it is in virtue of reason that every man 
presents his own existence to himself as an absolute end? 
Is not this a perverse way of regarding what is merely 
the animal instinct of self-preservation ? (2) Granting that 
every man in virtue of his rational nature does thus present 
his own existence to himself as an absolute end, how should 
there result from this a common ' objective ' end consisting 
in the rational nature of every man ? Would not the fact of 
every man being an absolute end to himself have just the 
opposite result — that of rendering a common object im- 
possible? (3) If, in order to get over this difficulty, you 
suppose that to present one's own existence to oneself as an 
absolute end is to present one's rational nature as such an 

} Grundlegung, &c. pp 276 and 277 ; Tr„ pp. 65 and 67. 


end, the question arises, (a) how reason can present itself 
to itself as an end, how it can be at once an absolute end, 
and the will which seeks that end ? (b) how, if it be so, vice, 
or selfishness, is to be explained, which would seem to imply 
the presentation by reason (as that in virtue of which a man 
makes himself his own object) to itself of an irrational object? 

113. By an instinct we mean a tendency, of the end or 
object of which there is no conception on the part of the 
subject of the tendency. From the nature of the case the 
analysis of our moral experience cannot carry us back to any 
such tendency, which, just on account of there being no con- 
ception of its end on the part of the subject of it, would have 
no moral character. It is the necessary condition of a 
moral act that it be determined by the conception of an end 
for which it is done. Hence actions done in sleep (resulting 
from animal tendencies), or strictly under compulsion, or 
from accident, have no moral character. We are apt, indeed, 
to speak of actions, which are in truth morally imputable, as 
if they were due to mere force, e.g. actions done to escape 
imminent danger, as when a man leaves the post of duty to 
save his life. But, in fact, such actions are determined by a 
conception of oneself as liable to a threatened pain, to avoid 
which the action is done. They are thus not the result of 
the animal instinct of self-preservation (of a tendency of the 
end or object of which there is no conception on the part of 
the subject of the tendency), but of a conception of himself 
as an end on the part of the agent, and just for that reason 
they have a moral character. In short, in order to become a 
spring of moral action (an action morally imputable, or for 
which the agent is accountable, an action to which praise or 
blame are appropriate), the animal desire or aversion must 
have taken a new character from self-consciousness, from the 
presentation of oneself as an object, so as to become a desire 
or aversion for a conceived state of oneself, or for an object 
determined by relation to oneself. It is because the moral 
agent is thus conscious of himself as making the motive to 
his act, that he imputes it to himself, recognises himself as 
accountable for it, and ascribes a like accountability to other 
men, with whom he could not communicate unless they had 
a like consciousness with his own. 

114. But admitting the presentation of the agent's self to 
himself as the object of his action to be the condition of its 


having a moral character for good or evil, we may ask why 
such 'presentation, &c. ' is to be ascribed to ' reason ' ? What 
has it in common with ' reason ' in that sense in which we 
suppose it to be employed in our knowledge of the relations 
of things and in the process of arriving at general truths ? 
The answer is that it is only through the consciousness of 
self that we are conscious of objects as related to each other 
in one world, and that thus self- consciousness is the basis of 
all our knowledge. But while there is this real identity be- 
tween the spirit of man as knowing and the spirit of man as 
morally acting (an identity properly enough expressed by 
speaking of reason as having a speculative and a practical 
employment), we must not ignore the equally real difference 
between the two employments. Mere abstract self-conscious- 
ness does not constitute either one or the other. In regard 
to knowledge, it is a false abstraction to separate reason 
from sensuous perception, as if the mere' senses gave us 
certain reports, and then a self-conscious principle wrought 
these into a connected system. ' Sense without reason is 
blind, reason without sense is empty.' Mere sense tells us 
nothing. Unrelated objects (and, apart from self-conscious- 
ness, objects would be unrelated) are no objects at all. On 
the other hand, consciousness of self, save as determined in 
relation to objects in sensuous perception, would be con- 
sciousness of nothing. Thus reason and sense are two sides 
of the one reality, knowledge. A like truth holds in regard 
to ' practical reason. 5 Mere reason, mere consciousness of 
self, apart from desire, would be no principle of action at 
all, for it would not be a presentation of a state of oneself, 
or of an object related to oneself, as a state or object to 
be attained or realised. Reason, then, in the sense of self- 
consciousness, is the condition of moral activity, inasmuch 
as the motive to such activity is the presentation of a state 
of oneself, or of an object related to oneself, as a state or 
object to be attained or realised. Such a motive may be 
described as desiring or wanting self -consciousness, or as 
self-conscious desire (vovs bpsKracos or SiavorjTt/crj ope%is). 

115. Such a motive issuing in act, or such an activity as 
so determined, = will, which thus, alike whether good or 
bad, necessarily involves practical reason; though there may 
be (as we shall see) a further sense in which only the good 
will can be identified with practical reason. This account 


of the will corresponds in effect to what may be called the 
generic notion of the will in Kant, though in his later moral 
writings he uses ' Wille ' in the specific sense of good will, 
(i.e. for desire determined not merely by self-consciousness, 
by any kind of conception of self as an absolute end, but by 
a true conception of self as an absolute end ; in other words, 
for will/ree in the sense of being autonomous, as opposed to 
that sense of freedom in which even the heteronomous will 
is free), while he uses 'Willkuhr' for the will free or 
rational only in that sense in which it must be so to be a 
spring of action morally imputable, good or bad. The 
account above given of will applies to it equally in the two 
forms or states which Kant in his later ethical writings (not 
in the Grundlegung) came to distinguish as 'Wille' and 
< Willkiihr.' 

Of will in the generic sense Kant gives in different places 
three definitions. It is (1) the faculty of bringing into exist- 
ence, or of setting oneself to bring into existence, objects 
corresponding to ideas (' Yorstellungen ') ; (2) the faculty of 
acting according to the consciousness (' Vorstellung ') of 
laws ; (3) a species of causality which belongs to living 
beings so far as they are rational. Freedom is a property 
which belongs to their causality so far as it can operate 
independently of determination by alien (or external) causes 
(' fremde Ursachen,' causes foreign to itself), while natural 
necessity is the property which belongs to the causality of 
all irrational agents, the property of being determined to 
activity by the influence of alien causes. 

116. As to these three definitions, observe, first, that they 
all imply an essentially different notion of will from that held 
by the psychologists (e.g. Bain), who regard it simply as 
an 'activity directed by our feeling.' Kant's definitions (1) 
and (2), indeed, would not be materially altered by the 
substitution of the term ' activity ' for * faculty ,' and 'faculty* 
is always a misleading term. In speaking of will as a faculty 
we are apt to convey the notion of its being other alike than 
man and his activity, something which belongs to him but is 
not himself. The truer way of thinking of it is as the man 
in relation to a certain sort of activity, or a certain sort of 
activity considered in relation to the man from whom it 
proceeds (as distinct from the consideration of it in relation 
to its effects). But then we have to ask, What sort of 


activity ? and here the difference between Kant's view of it and 
Bain's is essential. With one it is activity determined by 
feeling, with the other activity determined by reason. It 
does not, of course, matter intrinsically whether the term 
6 will ' is or is not applied to ' activity directed by feeling ' 
generally (in which sense it will cover the same ground as 
Aristotle's i/covcrcov, and like it be equally predicable of 
animals), or whether it is restricted to activity proceeding 
from desire and aversion of a rational subject, i.e. a subject 
conscious of himself as an absolute end. What does matter 
is the distinction between the two forms of activity, which 
is conveniently marked by keeping the term c volition ' for 
one, and c will ' for the other. 

117. Observe next that Kant does not himself say that 
by ' rational ' in definition (3) he means ' self-conscious ' in the 
sense explained. We are warranted, however, in saying (a) 
that Kant always understood self-consciousness to be in- 
volved in reason, though reason, as he often uses the term, 
means more than this; (b) that by understanding 'rational' 
as = self-conscious in definition (3), we are able to adjust 
this definition to the other two ; for the consciousness of 
self is the condition of there being a consciousness of objects 
which one may set oneself to bring into existence (mere 
feeling not being a consciousness of objects), and also of 
there being a consciousness of law; (c) that only by so 
understanding ' rational ' does the definition become equally 
applicable to the vicious and virtuous will. If Kant had 
been asked what he meant by ' rational ' in this definition, 
he would probably have said f capable of being determined 
by the consciousness of law ' ; and so far as rationality is 
understood to mean merely the capacity, as distinct from the 
actuality, of such determination, the definition will be equally 
applicable to the will as it exists in the morally good and 
the morally bad, as f autonomous' and 'heteronomous.' But 
self- consciousness is capacity for determination by the con- 
sciousness of law. It makes the difference between the 
natural agent determined according to laws and the moral 
agent capable of determination according to the consciousness 
of laws. Consciousness of law implies consciousness of a 
subject to which the law relates, and of this self-conscious- 
ness is the condition. Conversely, self-consciousness, the 
presentation of self as an end or as that to which all ends 


are relative, carries with it a distinction between that which 
is good as satisfying a present want, and that which is good 
for me on the whole; in other words, that capacity for 
determination by the conception of the desirable, as other 
than determination by desire, which may become determi- 
nation by the consciousness of law. Observe that anyhow 
' heteronomy ' or bondage of will does not = its naturali- 
sation ; it is not a condition like that in which one natural 
phenomenon is determined by 'fremde Ursachen,' because the 
determining object, whatever it is, is one which the agent 
makes his own (he is determined by his consciousness of 
his self as an absolute end). Kant, in anxiety to maintain 
that only will determined by rational laws is free in the 
sense of autonomous, is apt to write as if the heteronomous 
will were not a will, according to his definition of will, at 
all ; as if it were determined by ' fremde Ursachen ' in the 
same way as natural phenomena. But in truth, ' to admit 
a nature into the will ' in the sense of adopting a motive 
from animal susceptibilities (supposing that to be a true 
account of the bad will), is quite different from becoming 
natural, in that sense in which determination by ' fremde 
Ursachen ' is characteristic of the ' natural,' or from losing 
the characteristic of will as activity determined by the 
consciousness of self as an end. It is the nature of will to 
be free in this sense. Except as representing will thus free, 
action is neither moral nor immoral. 

118. The will, then, being equally as virtuous and as 
vicious a 'species of causality which belongs to living 
beings only as rational,' i.e. only so far as the living being 
is self-conscious, and thus presents himself to himself as an 
absolute end (in all his desire desires some conceived state of 
himself or some object determined in thought by relation to 
himself), the question arises how the good will can be dis- 
tinguished, as the will of which the object is man as an 
absolute end, from the vicious will of which the object is 
something else ; or as the will which treats humanity in the 
person of the agent and in all other persons only as an end ; 
or as the will which is autonomous, because determined by a 
law which itself as reason gives to itself, from the will which 
is heteronomous, as being determined by qualities of some 
object alien to itself, i.e. by anticipation of pleasure. Or, as 
the question was put above (§ 112), 'Granting that every man 


in virtue of his rational nature does thus present his own 
existence to himself as an absolute end, how should there 
result from this a common e objective ' end, consisting in the 
rational nature of every man ? ' 

In answering this question we must bear in mind that 
the presentation of self as an absolute end, which is involved 
in, and forms the rational element in, all willing and moral 
action, whether good or bad, is not a presentation of it as 
an empty and abstract self, but as a determinate self, as in 
a certain state determined by relation to certain objects, or 
of those objects as determined by relation to it; and, further, 
that the state and the objects have yet to be attained or 
brought into existence. (The relation of consciousness to a 
state or objects contemplated as already attained or actually 
in existence is not one of willing at all.) The character of 
the will, then, though it is always a presentation by the 
agent of himself as an absolute end, will vary according to 
the state of himself, or according to the objects determined 
in thought by relation to himself, which he seeks to attain, 
or, as we commonly express it, according to the nature of 
the man's dominant interests. The conceived object, to 
which in willing he seeks to give reality, may be a state of 
himself as enjoying certain animal pleasures, or a state of 
himself as fulfilling some vocation dimly conceived as be- 
longing to him in a divine plan of the world in virtue of the 
possibilities of improvement which he finds in himself. Or 
it may be (and more probably is, most men being neither 
sots nor heroic philosophers), some state of himself as filling 
a certain position in relation to his family or neighbours or 
fellow-citizens and finding happiness therein. Or it may be 
an object which could not naturally be described as a state 
of himself at all, but which is still determined by the relation 
in which he places it to himself, the ruin of an enemy, the 
happiness of a beloved person, the success of a political 
movement, the painting of a picture, the writing of a book, 
the improvement of his neighbours, the conversion of the 
heathen. There is thus great variety in the states under 
which, and the objects in relation to which, the self is pre- 
sented in that concrete presentation of it as an absolute end 
which is involved in willing, and the question is, whether 
they admit of being classified in two orders ; so that in so 
far as his will is directed to one order of states or objects, a 


man may be considered (a) to be living for an object common 
to himself with all rational beings and consisting in the 
perfection of the rational nature ; (b) to be living for 
humanity, c treating it in his own person and in that of all 
others as an absolute end,' and thus (c) to have his will 
' autonomous ' ; whereas, so far as his will is directed to the 
other order of states or objects, he may be considered to be 
(a) living for an object private to himself, separating him 
from instead of uniting him to other men and rational 
beings ; (6) treating humanity, the rational nature, in him- 
self and others as a mere means; and (c).to have his will 
6 heteronomous.' 

119. Perhaps there may be truth in such a classification ; 
but before we consider in what sense it is true, we must 
explain in what sense it is inadmissible. It is not true, as 
Kant seems to hold, that human motives are reducible either 
to desire for pleasure on the one side (in which case the will 
is ' heteronomous '), or desire for fulfilment of the moral law 
on the other (in which case alone, according to him, it is 
' autonomous ') . He is quite right in holding that no ground 
for distinction between higher and lower desires can be 
found in the exciting causes of the desired pleasure, if 
pleasure is in each case the object of desire ; he is quite 
right in holding that desire for pleasure, as such, is, from the 
moral point of view, all of a sort, and cannot be the motive 
of a good will. His error lies in supposing that there is no 
alternative between the determination of desire by anticipa- 
tion of pleasure and its determination by the conception of 
a moral law. It is this which leads him to say that c if my 
will is determined to act for the happiness of others, merely 
as for the attainment of an object of desire, it is really my 
own happiness (pleasure) that is the ground of the de- 
termination: nor is there any alternative between making 
this the ground of determination and finding it in the mere 
form of the maxim (' Seek the happiness of others ') as fit for 
universal law.' Now, so far as by this he merely means to 
protest against the notion that any man's obligation to seek 
the happiness of others is founded upon his desire for it, 
he is quite right. Such a notion can only mean that I ought 
to seek the happiness of others for the sake of the pleasure 
arising upon the satisfaction of the desire which I feel for 
this happiness. But it is quite another thing to suppose (as 


Kant seems to suppose) that a man's desire for the happiness 
of others must be ultimately a desire for the pleasure which 
he shall himself find in their happiness, unless it is a desire 
determined by the conception of a law commanding me to 
seek the happiness of others. In the ordinary concrete form 
of a desire to make this or that person happy, it is neither 
one nor the other, any more than is any other desire for an 
object. The notion, indeed, that the sole object of desire is 
pleasure, involves a confusion between the object which 
excites desire and the pleasure which ensues upon the satis- 
faction of the desire ; in other words, the mistake of sup- 
posing that desire is excited by the anticipation of its own 
satisfaction ; whereas, obviously, desire must previously have 
been excited by some other object before any such satisfac- 
tion can be anticipated. We are thus falling into a false 
antithesis, if, having admitted (what is true) that the 
presentation of self as an end conditions, or the quest of self- 
satisfaction is the form of, all moral activity, we allow no 
alternative between a motive consisting in the presentation 
of self as a subject to be pleased and that consisting in the 
presentation of it as a subject of law. Ordinary motives are 
neither of the one kind nor of the other. They represent 
interest in the attainment of objects without which the man 
cannot satisfy himself, and in attaining which he will find 
pleasure, but only because he has previously desired them. 
It is the object that he desires, not the pleasure which, 
having desired the object, he finds in the satisfaction of the 
desire. Such interests, though not mere appetites, because 
conditioned by self-consciousness, correspond to them as not 
having pleasures for their object. 

120. [A passage follows dealing with the relation of 
pleasure to desire, which is substantially embodied in the 
Prolegomena to Ethics, and does not bear directly upon 
Kant's doctrine. It concludes as follows.] 

The desires of men, then, as distinct from animal appe- 
tites, may be classified as (1) desires for pleasures incidental 
to the satisfaction of appetites ; (2) desires for pleasures 
other than these, for pleasures of pure emotion, and for the 
pleasures attendant on the satisfaction of interest ; this 
last form of desire for pleasure presupposes another kind of 
desire, which is not a desire for pleasure at all, viz. (3) 
desires for the attainment of sundry objects or for the reali- 


sation of sundry ideas, which may come to be followed, and 
perhaps sustained, by an anticipation of the pleasure which 
will ensue upon their satisfaction, but which must have 
arisen antecedently to this anticipation, and must be inde- 
pendent of any desire for pleasure. What, then, is to be 
said of the moral value of the desires thus classified ? 
According to Kant, only the interest which reason consti- 
tutes is morally good or the spring of a morally good action, 
and there is no such interest but that in the fulfilment of a 
universally binding and self-imposed law as such. All other 
interests, according to him, are reducible to desire for 
pleasure, and that is an interest with the constitution of 
which reason has nothing to do. It is as animals or parts 
of nature that we are susceptible of pleasure. Thus in 
being determined by pleasure as its object the will is 
c heteronomous,' determined by a natural influence which it 
admits into itself, but which does not proceed from it. Now 
we have tried to show that many of our desires, which 
involve no interest in the fulfilment of the moral law, are 
not desires for any sort of pleasure. We might fairly, too, 
reject the notion (if that were what Kant meant) that there 
is anything morally bad in the desire for c innocent pleasure * 
(desire, e.g. to repeat the pleasure which we experienced in 
hearing a piece of music yesterday). Does, then, Kant's 
distinction of the interest which reason constitutes, and in 
determination by which the will is autonomous and good, 
from those interests which the rational will (as self-con- 
sciousness) merely adopts and finds means to satisfy, thus 
suffering itself to be ' heteronomous,' afford any real criterion 
of the moral value of our motives ? 

121. I think that it does; but in order to put the question 
on a right footing, we must bear in mind that the proper 
subject of moral valuation is not this or that desire, but a 
character or action considered in relation to character. 
Desires, classified as above, in abstraction from the character 
as a whole of the man experiencing them, and from the 
effects on the character of the actions moved by them, are 
neither morally good nor morally bad. To desire pleasures, 
this, that, or the other, is one thing ; it is quite another to 
live for pleasure, to be a pleasure-seeker, to be a man with 
whom the desire for pleasure (not for ? pleasure in general,' 
for that is a fiction of hedonist philosophers, but for certain 


pleasures which the particular person finds specially pleasant) 
is the dominant interest. And though it would be a mis- 
take so to separate pleasures from their exciting causes as 
to consider them all of a sort (nor did Kant, I think, mean 
so to regard them), it is not a mistake to regard the living 
for pleasure (the character of the pleasure-seeker) as a moral 
condition which is the same whatever the differences of the 
pleasures sought by different pleasure- seekers in relation to 
their exciting causes may be : nor is it a mistake to regard 
it as a form, though not the sole form, of moral badness. 
The ground, too, for so regarding it is that the man so living 
(as will be explained) has for his dominant interest one 
which, though affected by reason like every other possible 
human interest, does not issue from reason as the motive 
power in man's development. Kant is only wrong in re- 
garding it (or seeming to regard it) (a) as a state in which 
certain men are living, whereas it is only a state to which 
they tend, an ideal 1 rather than an actual condition ; (b) as 
the only alternative to the state in which the sole interest is 
in the moral law. 

122. What, however, it may be asked, is the ground of 
distinction between desires for pleasures, which we say are 
morally neither good nor bad, and the living for pleasure as 
a mode of character and a mode which is bad ? Is not the 
living for pleasure equivalent simply to the series of desires 
for this or that pleasure, so far as these are stronger than 
other desires ? Is not the pleasure-seeker the man who is 
always desiring some pleasure or other more strongly than 
he desires anything else, and who thus acts from such desire 
(the will being simply a strongest desire) 9 With what 
meaning, then, can we ascribe moral goodness or badness to 
character, and deny it to desires which are simply character 
in detail ? 

By character, as that to which moral predicates are 
ultimately relative, we mean the way in which a man seeks 
self-satisfaction. The will is always an effort after self- 
satisfaction (as explained above, a presentation of some state 
of oneself or of an object determined in thought by relation 
to oneself, as to be attained or realised in preference to any 
other), and character depends on the direction which this effort 
takes, on the nature of the state or objects in which this self- 

1 [« Ideal ' is queried in the MS.] 


satisfaction is sought. No doubt it may sometimes be 
sought in one sort of object, sometimes in another, but the 
effort always tends to fix itself in a certain direction ; and only 
so far as it does so, is there a character which we can call good 
or bad. Thus to children, because in them this tendency is 
not yet fixed or ascertainable, we are only able to ascribe a 
moral character in some anticipatory way. It is not intended, 
then, to deny that desires as they actually are found in a 
man admit of being morally good or bad. What is meant 
is, that they only admit of this in virtue of the relation in 
which they stand to character, defined as above. E.g. a 
passing desire for revenge against a person who has insulted 
one, simply in itself, is neither good nor bad. The man who 
experiences it is (or tends to be) good or bad according to 
the mode in which the set of his character (the nature of his 
dominant interests, of the objects in which he has come to 
seek self-satisfaction) leads him to deal with it ; whether to 
keep out of the other man's way and distract himself with 
occupations till the desire has passed away, or to seek occa- 
sion to gratify it. Conversely with a desire to do a kindness 
to some one. This is not morally good in itself. It may be 
bad in relation to the character of a self-indulgent man, in 
the same sense in which desire for revenge may be good in 
relation to the character of a man whose dominant interest 
in good work so absorbs it, that it passes off in the shape of 
increased warmth in fighting some public nuisance. But a 
more exact way of speaking is to say that the man who has 
it is good or bad according to the direction which the set of 
his character gives to its manifestation. Kant in one place 
seems to say that benevolence is of no moral value if you 
have any natural liking for the persons towards whom you 
exercise it. Critics make merry over this, but it is true that 
benevolence only has a moral value so far as it belongs to 
the character of a man who has come to seek his self- 
satisfaction in pursuits which will make him do good to 
others whether he likes them or no. So in regard to plea- 
sure. The wish for this or that pleasure, and the act 
determined by it, is not bad unless either (1) the pleasure is 
one which would not be pleasant to a man who was seeking 
to satisfy himself with worthy objects, or (2) the wish for this 
pleasure belongs to a character which seeks satisfaction in 
pleasure, a character in which the quest of pleasure is the 


dominant interest. Some people (as we have seen) have 
come to fancy this the sole object of pursuit, because they 
necessarily seek self-satisfaction, and self-satisfaction when 
attained is pleasant. But the self-satisfaction which a man 
seeks, though ipso facto pleasant supposing it to be ever 
attained (pleasant with the pleasure which attends on every 
effort in the attainment of its end), is not properly said to be 
pleasure, unless pleasure forms the object with which in the 
general tenour of his life he seeks to satisfy himself. In 
that case we speak of him as the ( mere voluptuary.' 

123. While it is true, however, that pleasure- seeking, 
i.e. the set of the character towards pleasure, in whatever 
way excited, as that in which the man seeks his self-satisfac- 
tion, is essentially immoral, it is not the only type of im- 
morality. To consider it so is as much a mistake as to 
consider the conscious interest in the realisation of the moral 
law the only type of moral goodness. Immorality is selfish- 
ness, i.e. the direction of a man's dominant interest to an 
object private to himself, a good in which others cannot 
share. The character of the pleasure- seeker is necessarily 
selfish in this sense, but so are other forms of character. 
That the pleasure-seeker necessarily lives for an object pri- 
vate to himself may seem inconsistent with the fact that we 
c share each other's pleasure,' but it is not so. When a man 
is said to share another's pleasure, what is meant is that, 
having desired the same object with the other, he is equally 
pleased in its attainment ; or that, the pleasure of the other 
having been his object, he is satisfied when that object is 
attained, when the other is pleased. In each case the 
pleasure is private to the person enjoying it ; and so it 
always must be, even when it is incidental to the attainment 
of an object that is really common. It is only because we 
confuse the pursuit of a common object, i.e. of a good by 
which others than the pursuer will be the better, with the 
pursuit of the pleasure which will ensue when the object is 
attained, and thus regard those as pleasure- seekers who are 
not really so, that we come to imagine there can be pleasure- 
seekers who are not selfish, not living for an object purely 
private to themselves. But though pleasure-seeking is 
necessarily selfish and immoral, it is not the only form of 
immorality, considered as selfishness. A man may be living 
for an object other than pleasure, e.g. his own glory, and yet 


be purely selfish, inasmuch as his good (the object in which 
he seeks self-satisfaction), though not pleasure, is yet one 
in which others do not share, for which they are not the 
better. On the other hand, a man may be living for objects 
in the effort after which he takes no positive thought for the 
good of others, without being therefore selfish. An artist 
or man of science, who * lives for his work ? without troubling 
himself with philanthropy, is yet not living for an object 
merely private to himself. His special interest may be 
shared by no one, but the work which results from it, the 
machine constructed, the picture painted, the minute step 
forward in knowledge, i.e. the man's good as attained, is a 
good for which others are the better (whereas no one is the 
better for the fact that you or I are pleased now and again). 

124. It may seem that we have travelled a long way from 
the promised l vindication of Kant. This was to involve the 
reduction of all virtuous motives to manifestations of an 
autonomous will, as a will of which the object is (a) the ful- 
filment of a universally binding and self-imposed law, and 
(b) man as an absolute end. We have spoken of the good 
will, in relation to which alone any desire or act is morally 
good, as the will set in the direction of attaining worthy 
objects, and objects for the attainment of which others than 
the person willing them will be the better : of the bad will 
(of which the pleasure-seeking will is a conspicuous but not 
the sole type) as the opposite of this. But we have not 
shown how we are to estimate the worthiness of objects, or 
how in being determined by them the will is autonomous. 
To do so we must make a new beginning. 

[The passage which follows is substantially embodied in 
the Prolegomena to Ethics (see e.g. §§ 118 ff., 174-177, 193, 
200-205). The following is an outline of the argument : — 

6 Owing to the fact that man not only wants, but is 
conscious of himself as Wanting, conscious of himself as 
other than his want, though related to it, at once deter- 
mining and determined by it, there may and in some 
measure always must supervene upon his desire for this or 
that object, a desire for a more perfect state of himself. 5 
The conception of more fully realised possibilities involved in 
such desire may be called a conception of * a desirable as 
distinct from the desired.' ' The capacity for desiring an 

1 [Section 120.] 
VOL. II. £ 


object because thus conceived as desirable ... is the foun- 
dation of man's moral nature.' 1 The moral value of a 
character . . . depends on the degree to which such desire 
is habitually dominant.' Now the self, the fuller satisfaction 
of which is presented as thus absolutely desirable, is from 
the first a self ' existing in manifold relations to nature and 
other persons,' and ' these relations form the reality of the 
self.' Thus the conception of a self to be satisfied neces- 
sarily carries with it the conception of this object 'as 
common to himself with others.' Such a conception in its 
most primitive form is the germ of what Kant calls a * Eeich 
der Zwecke,' a ' kingdom of ends.' 

Thus the object presented to itself by reason as the 
absolutely desirable is from the first potentially what Kant 
requires of the object of the good will, viz. (a) an object for 
all rational beings, (b) the source of a law binding on all 
rational beings as such. It is (a) because, ' so far as A and 
B are each interested in an ideal of his own perfection, the 
object of their interests is really the same,' and (b) because 
' the man who is conscious of his own better being ' as his 
object, is ' conscious of a principle of action which from its 
nature, whether he acts from it or no, is universally and 
unconditionally applicable to his life, and which at the same 
time conflicts with motives that cannot be acted upon if it 
is to be acted upon.' Such an object need not be adequately 
conceived as the common good, but only ' up to the lights ' 
of the individual in question : on the other hand, it does 
need to be recognised by him as giving rise to a law 
unconditionally binding on him. No particular object or 
law can be presented which completely satisfies these require- 
ments, except the object which i consists in the disposition to 
seek perfection,' and the law * be perfect.' But nevertheless 
6 an object of merely relative value may be pursued with a 
devotion which arises from a consciousness, unable clearly 
to interpret itself, that there is something which has absolute 
value, and a law of limited validity may be obeyed with 
loyalty due to an assurance of there being an universal law, 
which cannot state to itself with adequate universality what 
that law enjoins.'] 

125. So far our concern has been to show that the 
principle of good character, being desire determined by the 
conception of the absolutely desirable, is for that reason a 


prevailing readiness to conform to a universally binding law 
of conduct because recognised as such. The rule of conduct 
upon which the good man acts is not always, we find, 
(perhaps never) one which, in any form in which he could 
state it, is fit to be universal law ; but to him it bears an 
authority derived from an ideal of absolute good, of which 
the operation upon him transcends his powers of definite 
intuition and expression, and is therefore presented as having 
a claim upon his obedience not conditional upon his likes 
and dislikes, a claim in that sense unconditional and uni- 
versal. His goodness consists in his practical recognition 
of that claim. But it is a further point in the Kantian 
doctrine, that in order to be good a man must conform to a 
universally binding law, not merely as universal, but as self- 
imposed. Upon this two questions arise ; (a) In what sense 
is the moral law in its truest form really self-imposed? 
(b) Is it necessary to moral goodness that the rules of con- 
duct which a man recognises as unconditionally binding 
upon him should also be recognised as self-imposed ? 

Answer to (a). The presentation of a moral law, which 
is the same thing as its existence, arises, as we have seen, 
out of the conception of the absolutely desirable, and this 
again arises out of the consciousness in man of himself as 
having the possibility of becoming something which he is 
not actually, but which he must become in order to be 
satisfied with himself. In other words, the moral law is the 
product of the individualising principle in man, that which 
alone enables him to say, ' I am myself and not another,' 
and to think of anything as his own. We properly enough 
represent this state of the case by saying that the moral law 
is self-imposed. This is quite compatible with saying that 
it is not of our own making in the sense that it is not the 
product of any desire or aversion, or of any number of desires 
or aversions, which any one of us or any number of us happen 
to have. It does not rest with you or me, in the ordinary 
sense of the words, or with anything which we may or may 
not will or do, whether there shall be such a law or no, any 
more than it rests with us whether we shall or shall not be 
rational being3, though it is through reason that each is a 
self in the only intelligible sense, i.e. as self-conscious. It 
is quite compatible also with saying that it is of divine 
origin. We rightly consider it so (rightly call it the divine 

L 2 


command or expression of the divine will) if we are right in 
holding (a) that the human self-consciousness is in principle 
identical with the divine — i.e. identical with it in respect of 
its form, or as self -consciousness, though not in respect of the 
limited matter which this self-consciousness takes into itself, 
or of its development in time ; and (b) that that perfection 
of persons in a perfect society, that c Eeich der Zwecke,' 
towards which we are perpetually struggling, but which, 
under the conditions of human life, can never be fully 
attained, exists as eternally complete in God. To those who 
so think, if the conception of morality as obedience to divine 
law comes to seem scarcely appropriate (because God cannot 
properly be thought of as standing to us either in the rela- 
tion of a political superior or in the relation in which desire 
determined by the conception of an absolute good stands to 
other desires as a restraining c imperative '), it is only to 
have its place taken by the conception of it as the fulfilment 
of a divine vocation, or as submission to a necessity, to 
which in one way or another all submit, in that way which 
makes the difference between the morally free and the 
morally ' bond,' viz. a willing submission from the recognition 
of it as a divine order which leads those who so recognise 
and submit to it towards that union with God in which one's 
own perfection is to be found. 

126. (b) But must the law of conduct be also recognised 
as self-imposed, in order to true goodness? To say so 
seems at first to contradict the 6 conscience ' of the un- 
sophisticated man. To him it seems that in doing his duty 
he is emphatically not his own master ; that it is imposed 
on him, as he would probably say, from without. In the 
perplexities of modern controversy he will often be found 
to protest that if there is no external imponent of moral law, 
there is no such law. What he means to convey by this, 
however, is that it must be what Kant calls * objective,' not 
dependent on the individual's likes and dislikes, not a 
product of any desire other than that determined by the 
conception of the absolutely desirable : but he interprets to 
himself the * objective ' under the figure of the * external,' 
an interpretation of it which must clearly be understood 
in some very loose way, for strictly the term 'external' 
expresses a relation in space, a relation which cannot obtain 
between the conscience and the law which it recognises. 


Granting, however, that the conception of the moral law as 
' objective ' in the sense explained, as dependent not on any 
principle of action private to the individual himself (such as 
desire for this or that pleasure, or for any object which is an 
object to him in virtue of his special temper and tastes), but 
on the consciousness of an unconditional good common to 
all rational beings, satisfies all the requirements of con- 
science ; granting that the plain good man only demurs to the 
notion of the moral law being self-imposed, and insists on its 
being something ' external ' to him, because he does not quite 
understand what is meant by ' self-imposed ' and e external ' ; 
still it is another matter to say that it is necessary to the 
goodness of the good man that he should conceive himself 
as the author of the law which he obeys. On consideration 
we shall see that the self-imposedness of the moral law is 
really implied in the absoluteness or finality of the moral 
object. If a man is not thoroughly good whose act is 
' pflicht-massig ' (' conformable to duty '), but not done ' aus 
Pflicht' ('from duty '), whose outwardly virtuous act is done 
either under force or from fear of punishment or hope of 
reward, then a man is not thoroughly good whose motive to 
right action is derived from anything else than his own con- 
sciousness of the absolute value of right action. Such a man 
will tell us that he does whatever good act he does because it 
is the will of God, or the will of his king or country, or the 
will of his parents, &c. or what his neighbours expect. But 
he is also conscious that he does not serve God or king, &c. 
aright, if he does it for any ulterior motive, for hope of 
reward, or fear of punishment. Why, then, does he do it ? 
Because he conceives that he ought, cannot bear to think of 
himself as not doing it, sees something intrinsically desir- 
able (viz. his own perfection) in it. The 'sic volo sic 
jubeo ' is thus the ground of his action. The law which he 
obeys is really self-imposed, and he is conscious of it as 
such, though the consciousness requires a certain ' maieutic ' 
before it can be brought to the birth in a shape in which he 
will acknowledge it, and the delivery is impeded by the 
difficulty of distinguishing himself, as the author of the law 
which he obeys, from those desires which are his own and 
which he imputes to himself, but which conflict with obedi- 
ence to the law. It is not necessary to his goodness that 
this ' maieutic ' should have been fully performed, and that 


lie should be prepared to regard himself as the author of the 
law which he obeys. What is necessary is that he should 
be independent of all inducements to obey it beyond the con- 
sciousness of the obedience itself as a thing absolutely good, 
an end in itself, which implies, whether he admits it or no, 
that it is self-imposed. 



[See Critique of Practical Reason, first part, I. 1, § 3, Eemark II.] 

127. ' All men, according to a natural law, seek pleasure. 
The way to find the moral law or rule of conduct is to find 
how this end, necessarily sought, may best be attained.' 
To this Kant objects that if the uniform means of attaining 
the maximum of pleasure could be discovered, the resulting 
rule of conduct would have no ' objective necessity,' but only 
a necessity of the same kind as the necessity of desiring 
pleasure, i.e. a physical necessity, which from the moral 
point of view is an accident depending on subjective ! suscepti- 
bility to pleasure. In order to understand Kant's view, we 
must distinguish man's relation to the physical world, as a 
phenomenon among other phenomena, from that relation to 
himself in respect of which he is a moral agent. His sus- 
ceptibility to pleasure and pain is a physical relation. If 
the imagination of pleasure and pain were uniform deter- 
minants of the animal faculty of desire, this determination 
would be an ' objective law of nature, 9 not a moral law at all. 
Looking at action, as part of nature, from without, this law 
would be c necessary ' for us with the ' necessity ' which 
belongs to our conception of nature, which compels us to re- 
gard phenomena as a system. This necessity, indeed, is in 
one sense according to Kant of subjective origin. It results 
from the presence to all phenomena of that which is not 
itself a phenomenon, viz. the ego; but the result is the 
* objective necessity of natural laws,' laws which, but for the 

1 The terms 'objective' and 'sub- of the moral law, and which is thus the 
jective' shift their meaning according same for all rational agents, the 'sub- 
to the point of view. ' Objective ' ject ' to which it is opposed being the 
means that which does not depend on individual as susceptible of pleasure, 
the ' subject,' and in each case we have Taking the ' subject ' as the rational 
to inquire what the ' subject' is from self, then the moral law would be ' sub- 
opposition to which its meaning is de- jective' as being a law which this sub- 
rived. Here ' objective ' means that ject imposes on itself, 
which depends on reason, as the source 


ego, would not be, but are not laws of the ego, do not apply 
to it. Moral laws, on the contrary, if there are such, are 
not laws of relations of phsenomena inter se — not laws merely 
resulting from the relation of the ego to phsenomena — but 
laws imposed by the ego on itself, and regulating, not the 
interaction of phsenomena, but the relation of the ego to 
phsenomena, as a subject formally self-conscious and self- 
determined, but only gradually realising its self-consciousness. 
That alone has necessity as moral law, or for man as a moral 
subject, which arises out of the relation of reason to itself as 
its own end, as striving to give reality to itself — to its own 
possibility — in the phsenomenal world. In contrast with this 
necessity the relation of imagined pleasure to animal desire 
is a mere accident. Its physical necessity lies outside his 
moral life. On the other hand, the relation of conceived 
pleasure to the rational impulse after self-satisfaction is an 
element in his moral life, for in selfishness it constitutes the 
matter, the transformation of which is the problem of the 
moral life. 

128. It may be said, ' Why should not ' physical neces- 
sity ' suffice ? What * objective necessity ' in a law of action 
need we or can we seek but ' physical necessity ' ? Is not 
the supposed ' necessity ' of Kant's moral laws (the necessity 
represented by the judgment, 6 1 ought,' ' as a rational agent 
I must') itself properly a 'subjective' necessity? i.e. a 
necessity merely arising from the subject's habit of mind, 
the habit of conformity to laws representing the general 
convenience of society, to be accounted for as a tendency 
gradually defined by the mutual action and reaction of man 
and his environment, and transmitted ? ' The answer is, 
that physical necessity is necessity in the sequence and 
simultaneity of sensible events. The connection between the 
conception of an end or the conception of a law and an action 
is not a sequence of this sort, and the law regulating it, from 
the nature of the case, is not a physical law. A rule of con- 
duct, derived from observation of the physical consequences 
of actions in the way of producing pleasure, is still not a 
physical law. A merely natural agent cannot present a rule 
of conduct to itself. The presentation of it, as a rule of 
what should be done in distinction from what is done, arises 
from the effort of reason, as a principle of self-realisation, 
conditioning and conditioned by an animal nature, to 


become what, as so conditioning and conditioned, it is not ; 
the effort to find an end adequate to itself, which it can in 
truth only find by making it, by giving reality to its own 
possibilities. The ' good,' the ' desirable ' (as distinct from 
the desired), the *■ should be,' the ' moral law,' are different 
ways of expressing the relation of the self-conscious subject 
to such an end. So long as reason seeks it in what does not 
depend on itself, in what it finds but does not make, in 
pleasure, which is the satisfaction of an animal susceptibility, 
just as possible without reason (Kant would say, much more 
possible), it is seeking it in what relatively to it is accidental, 
in what does not arise out of the principles through which 
alone there comes to be a ' should be ' at all. To say, in 
short, that I ought to pursue an end, viz. pleasure, which 
(as those who say it ought to be pursued strongly insist) 
in virtue of my animal nature I inevitably do pursue, is 
absurd. Just because the pursuit of pleasure is a physical 
necessity (though not therefore a necessity to us who are not 
merely physical), it cannot be morally necessary — cannot be 
that which morally must be. 



129. The points of difficulty in Kant are, (1) The op- 
position of the idea of the moral law, which alone determines 
the 'good will,' to 6 Erfahrungsbegriffe.' If not derived 
from experience, what is its origin, and what is its content 9 
Is there any alternative between its being derived from ex- 
perience and its being innate 9 And is not the condition of 
children and primitive races, nay, of such a people as the 
Greeks, in whose philosophy it is a received commonplace 
to say that the idea of duty does not appear, fatal to the 
supposition of innateness ; to say nothing of Kant's own ad- 
mission, that observation suggests the doubt whether an act 
corresponding to his definition of a morally good act, as 
proceeding from the pure idea of duty, was ever done ? In 
like manner, is there any alternative between leaving the 
idea of duty a mere empty abstraction, an idea of nothing in 
particular to be done, and appealing to experience to tell us 
what our duty is 9 Does not Kant himself implicitly make 
such an appeal in adopting the rule that the maxim of an 
act must be one fit to be universal law, for how can anything 
but experience settle this fitness 9 l 

(2) The doctrine that no result of any kind can con- 
tribute to, or detract from, the moral goodness of an act. 
Will not this (a) justify some of the most mischievous acts 
that are done, which yet the agent does * conscientiously ' 9 
and (b) leave us without any sure standard by which to 
judge of the actions of others, and liable to much self-deceit 
in regard to our own, since everyone can flatter himself that 
his motive was good, however much mischief may have re- 
sulted from his action 9 2 

(3) We may admit that not the actual result, but the 
intended result or motive, gives the moral character to an 

1 [See above, sections 104-105.] 

2 [See Prolegomena to Ethics, book iv. ch. 1.] 


action. Does not Kant, however, make it the condition of a 
good act that every possible motive to it should be absent ? 
It is not to be determined by any ' motive a posteriori, 9 but 
only by a ' principle a priori,' the 'posteriority' and 'priority' 
being posteriority and priority to desire. An instance of a 
6 motive a posteriori, 9 which he himself gives, is natural 
benevolence, the desire to give pleasure to another. If a 
beneficent act is to be morally good, according to him, such 
desire must be absent. Thus it would seem (a), that what 
we are apt to think the best acts, acts of instinctive kind- 
ness, are set down as having no moral value ; and (b), that 
the morally good act remains without a motive, i.e. practi- 
cally impossible. Aidvoia avrrj ovdsv klvsI, A principle 
prior to all desire can be no principle of action at all. 1 

(4) What is meant by the ' objective necessity ' of the 
moral law and morally good action? How can action be 
* objectively necessary ' which scarcely anyone does ? Can 
' law ' be properly applied to anything but the command of a 
political superior ? If the application of the term can be ex- 
tended to any uniform order, as when we talk of ' natural 
law,' how can it be applied to that of which you can neither 
say that it is the command of a political superior, nor that 
it is a uniform order of phenomena ? How, if we allow our- 
selves to talk of moral law, can it be ' objectively necessary,' 
when scarce anyone conforms to it ? 2 

1 [See above, section 110-111, and 122.] 

2 [See above, section 125-126.] 



Note of the Editor. 

The following lectures on logic were delivered in 1874-75 when Green 
was a tutor at Balliol College. They take the form, not of a systematic 
exposition of the subject, but of a commentary and criticism on H. L. Mansel 
and J. S. Mill, the most representative of the writers on logic who were at 
that time studied in Oxford. In some points (especially in D and G) they 
will be found to supplement the preceding lectures on Kant. 

158 LOGIO. 




1. As to the office of logic, there are two principal views 
among modern writers, (a) that of the ' formal logicians,' 
of whom Hamilton and Mansel are specimens, (b) that of 
all others, the view, viz. that logic is the science of the 
method of knowledge. Mill, Kuno Fischer, Sigwart, Ueber- 
weg, under various expression, agree in this, though their 
views of what the method of knowledge is vary according to 
the difference in their notions of what the object of know- 
ledge is. This question, what the object of knowledge is (or, 
How is knowledge possible ? What are the presuppositions 
as to our relation to the objective world from which we must 
start in inquiring what the method is by which we come 
to scientific knowledge?), is the question of metaphysic, 
according to those who believe in the possibility of meta- 
physic. Those who do not would generally say that the 
questions with which metaphysic has professed to deal can 
only be dealt with by psychology, as the method of ascer- 
taining by experiment and observation how men reach the 
stage of consciousness in which scientific reasoning, or the 
establishment of truth upon evidence (of which logic analyses 
the method), becomes possible. Any way, whether we get 
our answer to the above questions by ' metaphysic ' or c psy- 
chology,' or by both, the answer must at least affect logic, 
according to the above account of it. 

There are those (a) who identify logic and metaphysic ; ■• 
who hold that the question of logic, What is the method 
by which knowledge is attained? is inseparable from the 
question of metaphysic, What are the necessary forms 
(the primary relations) of the objects of knowledge or the 
objective world ? There are those (ft who, admitting the 
possibility and necessity of metaphysic (as the science of the 


categories), hold that from the inquiry, What are the con- 
ditions under which an object that is to be known must 
exist? can be separated the inquiry into the process by 
which we, as individuals of imperfect and growing intelli- 
gence, attain to knowledge, a process which, as a whole, 
forms the subject-matter of psychology, and of which that 
part which we can consciously regulate (which it rests with 
us to perform correctly or incorrectly, and which can thus 
be corrected by us in virtue of the recognition of certain 
rules), is the subject-matter of logic as part of psychology. 
It may be held, e.g. that the question, How is it that the world 
exists as a quantity? (or, What is the origin of quantity?) 
is a metaphysical one, the inquiry into the general nature of 
quantitative reasoning a logical one : again, that the question, 
How is it that there is a system of nature ? (What is the 
origin of the 'principle of the uniformity of nature'?) is a 
metaphysical one, the analysis of the methods of induction 
(observation and experiment) founded on this ; principle, a 
logical one. According to this, though metaphysic and 
logic may be distinguished, metaphysical theory must con- 
dition logical. There are those finally (c) who discard 
metaphysic altogether, holding that \ nothing isTrequired to 
render reasoning possible but the senses and association ' ; 
that thus the presuppositions of knowledge, or conditions 
under which all knowable objects must exist, are, in fact, 
habits or tendencies of our minds which, by a process of 
sensitive experience, extending over innumerable generations, 
but ascertainable by experiment and observation, have be- 
come uniform or ' necessary,' and that the difference between 
this process and that "which logic investigates lies in the 
fact that the latter admits of being correctly or incorrectly 
performed according as it does or does not conform to rules 
of which we are conscious. Such persons must adjust their 
doctrine of scientific method to this doctrine, as, e.g. Mill 
tries to do by making out that the principle of the uniformity 
of nature, which the inductive methods presuppose, is the 
result of an unconscious induction constantly going on, — if 
this means (as it should to a Humist) that the above principle 
= a habit of expectation formed without our being aware of it. 
2. From the above view in all its forms is to be dis- 
tinguished that of ' formal logicians ' (in the Hamiltonian 
sense). With them logic is the science, not of the method 

160 LOGIC. 

of knowledge (which implies relation to objects), but of those 
c forms of thought ' in conforming to which we think cor- 
rectly, but in a way that contributes nothing to knowledge 
or truth. This view goes on the supposition that while, as 
all agree, knowledge has to do with real objects, there are 
processes of thought which do not affect and are not affected 
by such objects. It is inconsistent, then, alike with the 
doctrine that objects are only objects as for thought (and 
that thus every correctly performed process on the part of 
the thinking subject is a modification of the objective world, 
and every modification of the objective world is a further 
determination of the thinking subject), and with the doctrine 
that thoughts are merely a result (a symbolic summary) of 
sensible events. 

It derives its account of the formal processes of thought 
in the main from Aristotle, to whom, however, the opposition 
between correct and true thinking, between thought and 
knowledge, between laws according to which the subject 
thinks, and laws according to which objects are known, is 
wholly alien. Aristotle conceived his logic to be an account 
of the process by which the world is known, and this process 
to be a reproduction of the order in which the world exists. 
The process by which we know, as distinct from that by 
which we come to know, is one from (pvosc irporspa to $vctm 
varspa. The worst of it was that, as he had no clear concep- 
tion of any way of getting to the <pvasi, irpcora but that of 
successively subtracting from properties connected by a 
general name till a minimum of meaning was arrived at, so 
he had none of the progressive specification of the 'general 
idea (which is the true process of knowledge, corresponding 
to evolution in nature *), but that of passing from the name 
of least meaning, and therefore applicable to most things, to 
names of more meaning applicable to fewer things. Thus 
the Aristotelian or syllogistic logic earns the reproach of 
consisting in a series of verbal propositions. It represents 
neither a method of arriving at knowledge nor the system of 
ideas which constitutes the known world (in which general 
laws are so specified by particular conditions as to account 
for events), but is merely of use for analysing what is in- 
volved in conceded general propositions. Thus it is specially 
adapted for the purposes for which the Church, according to 

' [The words ' corresponding . . . nature ' are queried in the MS.] 


the Catholic theory, needs a logic. Hence its use by the 
Schoolmen. They did not want a method of arriving at 
truth, nor a theory of what knowledge consists in, for all 
truth was supposed to be conveyed, as the object of faith, 
not of knowledge, in revelation. What they did want was 
a method of evolving what was involved in conceded pro- 
positions of the faith. Nominalism is the process by which 
scholastic logic destroys itself. It is the recognition of the 
fact that in its deductions from universals syllogistic logic 
was merely analysing the meaning of names. Hence the 
modern mind, in the effort to know the truth about nature 
itself, discards it. Descartes, Spinoza, Leibnitz move in as 
complete freedom from it as Bacon or Newton. Practically 
it has continued in vogue as a method of arguing or dispu- 
tation (in distinction from a method of knowing or learning 
to know), specially of theological and forensic argument, 
for which it is well fitted, because, in common life, the object 
of argument is generally conviction of inconsistency (proof 
that an opponent ignores something involved in a proposition 
which he concedes), while the object of forensic and ordinary 
theological argument is of like sort, viz. to show that a 
general rule of law covers some particular case, or that a 
text in Scripture or a Father implies some particular doc- 
trine, which the author of the text probably never contem- 
plated, but which may be syllogistically deduced from it, 
because verbally covered by it. Applied to such purposes, 
— in other words, as a rule for securing consistency in the 
interpretation and application of general terms, — syllogistic 
logic has its value (a value as a practical, though not as a 
speculative, science). 

3. But meanwhile — chiefly upon the strength of the 
least true or valuable side of Kant's doctrine — a theory 
of formal logic has been constructed, which is incompatible 
even with the secondary office assigned to it above. Ac- 
cording to this view, it is a speculative science, which 
investigates the laws of 'formal 9 "thinking ; and 'formal' 
thinking means a process (a) of formal conception, by which, 
given certain attributes, they are conceived to represent one 
or more objects without reference to anything other than 
those attributes ; (b) of formal judgment, by which, given two 
concepts (attributes conceived as representing objects, as 
above), one attribute or set of attributes is thought as either 


162 LOGIC. 

contained* in, or excluded by, the other without reference 
to anything beyond these several attributes or sets of attri- 
butes; (c) of formal reasoning, by which, given two judg- 
ments having one term in common, and without reference to 
anything beside those judgments, an individual or a class is 
thought as contained in, or excluded from, a certain set of 
attributes or class, because contained in, or excluded from, 
one which that set or class contains or excludes. 1 The 
process in each case is governed by the 'principles of 
identity and contradiction.' In virtue of these, given cer- 
tain attributes, we by ' pure thinking ' (i.e. without reference 
to anything else), (1) conceive them, if contradictory, as not 
representing the same object, (2) conceive the object to 
which in any case they are referred as ' thereby limited and 
separated from all other objects, as being itself and nothing 
else.' 2 Again, given the concepts p and q, ' if q contain the 
attributes o, p, i, I can by a law of thought alone determine 
that all q is p, or, if q contain an attribute contradictory of p, 
I can in like manner determine that no q is p.' 3 The same 
applies to mediate judgment or syllogism. 

4. It may be asked, first, whether this account represents 
any process of thinking whatever, — whether the ostensible 
result of pure thinking is not exactly the same as its begin- 
ning. The strict formal logicians virtually admit that it 
represents no process in this sense. Thus in formal con- 
ception * attributes are given ' ; if so, an object or objects 
which they determine (i.e. distinguish from other objects) is 
also given, for otherwise they are not attributes. What is 
there then in the result achieved by pure thinking, in the 
form of conception (which is described as reference to an 
object), which is not explicitly in the data ? There are two 
ways in which it may be sought to meet this objection. 

(a) ' Formal conception ' may be taken to mean the mental 
act by which attributes are constituted, by which that which 
would otherwise be merely a feeling becomes a property of 
a feeling subject or of a felt thing. But so to understand 
it, however correct in itself, is inconsistent with the whole 
doctrine of formal logic, according to which formal thinking 
takes its materials as given, and neither adds to nor alters 
them. But ' formal conception,' according to the suggested 

1 See Mansel's Edition of Aldrich, Artis Logic® Rudimcnta, Introduction, pp. 
lxii-lxviii. 2 lb. p. lxv. 3 lb. p. lxvi. 


interpretation of it, would be an act which, if it could be 
fitly said to have a ' matter ' (feeling) given it at all, at any 
rate absolutely transforms it. Again, whereas formal logic 
turns on the separation of 'laws according to which the 
subject thinks ' from ' laws according to which the object 
is known,' the first act of formal thinking, if the above 
suggestion is adopted, would be one which constitutes the 
objective world and is the common principle of all know- 
ledge. Not only, then, will it not be admissible as a part of 
formal thinking ; the admission that there is such an act of 
thought at all is fatal to this doctrine of formal thinking, 
as showing that the antithesis between ' laws according to 
which the subject thinks ' and ' laws according to which the 
object is known ' is untenable. 

(Jb) It may be said that, though ' attributes are given 9 to 
formal conception, and with attributes necessarily objects, 
yet as given they are in a state in which contradictory 
attributes are liable to be referred to the same thing, and in 
which objects really identical, because thought under the 
same attributes, may yet be supposed different ; and formal 
conception gets rid of this confusion. This, however, is in- 
compatible with the strict 'formal' view. If certain given 
attributes are not explicitly recognised as contradictory to 
begin with, it is not by mere formal thinking that they are 
conceived as necessarily belonging to different objects, be- 
cause in that case reference to something other than the 
attributes, as already conceived, is needed in order that an 
incompatibility between them, not yet recognised, may 
become so. It does not follow that there must be 'new 
experience,' but, if not new experience, there must be some 
further consideration of connection between different ele- 
ments of what we already know, which implies that some- 
thing other than the * given attributes ' is taken into account. 
So, too, when we come to formal judgment: the concep- 
tion of the attributes o and p as included in q must be 
explicit, or it is not mere formal thinking that elicits 
them from q in the judgment ' all q is p,' &c. ; and the con- 
ception of inclusion being thus explicit, there is absolutely 
nothing in the judgment that is not in the conception, and 
there is no process of thought from one to the other. The 
same holds, mutatis mutandis, of formal inference. 

5. The formal logicians do not disguise this. To repre- 

M 2 

164 LOGIC. 

sent attributes which formal thinking finds to be not 
referable to the same object, they take ' a and not a,' ' white 
and not white/ ' round and not round,' i.e. attributes already 
cleared from all possibility of confusion with each other. 
As c not round,' e.g. has no determination or qualities of its 
own except as the contradictory of 'round,' there can be no 
real act of thought involved in the transition from the pre- 
sentation of such attributes to the conception of them as 
necessarily belonging to different things. In like manner 
the formal logician, in order to make sure that the thinking 
involved in immediate and mediate inference shall be merely 
formal, seeks for such expression of judgments as removes 
all appearance of there being any thought involved in the 
inference at all. ' Quantification of the predicate ' is no 
doubt required by the principles of formal logic. But, when 
once judgments are expressed in quantified form, the 
formal inference from them is reduced to nil. So long as the 
judgment stood, * all men are mortal,' there was some 
colour for saying that in the judgment, ' some mortals are 
men,' there was a further act of thought : but put it as ' all 
men = some mortals,' and the conversion into ' some mortals 
=all men ' loses all appearance of forming a further act of 
thought at all. 

6. This brings us to the second main question that may 
be asked as to the value of formal logic. Granted that it 
does not represent any process of thought (from something 
previously thought to something not thought in thinking 
the former), still it may be held that it rightly represents a 
complete act of thought ; that the tendency of thought is to 
reach the form which formal logic represents, according to 
which an attribute is thought as contained in a group, and 
that again in a more complex collection, or, conversely, an 
individual in a class and that again in a larger class. Is 
this true? Is thought, in its complete form, a series of 
quanta, one including, or included in, or equal to, the other? 
This takes us back to the nature of logical conception and 
judgment. With the formal logician a concept is a notion 
of an attribute or bundle of attributes fixed by a sign, 
which in the act of conceiving I regard as representative of 
an individual ; or (as Hamilton seems generally to put it) a 
concept is a notion of a class, as possessing some common 
attribute or attributes, to which class in the act of conceiving 


I refer the individual. In logical judgment, either an 
attribute (predicate) is thought of as contained in a group of 
attributes (subject), or a smaller class (subject) as contained 
in a larger (predicate), according as one or other of the 
above views of the concept predominates. Meanwhile 
with the constitution of the individual object and its attri- 
butes, thought is supposed to have nothing to do. This is 
6 perceived ' or ' presented in intuition,' and from it (accord- 
ing to this theory), as thus from time to time presented, the 
mind has ' abstracted ' attributes, and put them together as 
a concept fixed by a name (or, in Locke's language, into 
a nominal essence), upon which, as things have been found 
to which the name was applicable, a class has been formed. 
The ordinary act of conception takes place when some per- 
ceived object (or ' phenomenon ') is ' found to agree with the 
abstract idea' (as Locke would say), ' is referred to the class ' 
(as Hamilton would say), ' is thought under the attributes 
included in the concept ' (as Mansel would say) : ' this is a 
man,' and so forth. Here, it would be said, there is a 
presentative element and a representative, the former indicated 
by the ' this,' the latter by the predicate, which shows that 
the presented object, or intuition, is referred to a class, or 
brought under a concept, representative of an indefinite 
number of other like phenomena. The difference of this 
from the formal act of conception would be that in the latter 
case, though an object of intuition (or individual object) is 
thought under the attributes, — in other words> though these 
are taken to represent such object,— it is a possible object of 
intuition, not an actual. (E.g. having a concept contain- 
ing the attributes of roundness, whiteness, softness, I 
conceive a round, soft, white something, — thus, in Mansel's 
language, 'individualising my concept,' — though I do not 
perceive anything.) 

7. ' Intuition,' according to Mansel (Prolegomena Logica*, 
p. 9, note), = ' every act of consciousness of which the imme- 
diate object is an individual, thing, state, or act of mind, 
presented under the condition of distinct existence in space 
or time.' It includes ' all the products of the perceptive 
(external or internal) and imaginative faculties ' (ib.). ' In 
every act of consciousness the ultimate object is an indi- 
vidual. But in intuition this object is presented to the mind 
directly, and does not imply the existence, past or present, 

166 LOGIC. 

of anything but itself and the mind to which it is presented. 
In thought, on the other hand, the individual is represented 
by means of a concept, which contains certain attributes 
applicable to other individuals of the same kind. This 
implies that there have been presented to the mind prior 
objects of intuition, originating the concept or general 
notion to which subsequent objects are referred. Hence 
arises another important distinction. All intuition is 
direct and presentative ; all thought is indirect and repre- 
sentative ' (ib. p. 11). 'Sensation' (apparently) is not yet 
' intuition ' : it is to intuition as mere feeling to felt thing, 
a feeling not attended to, not distinguished by 'this,' 'here,' 
or ' now.' l 

We must distinguish (1) sensation without conscious- 
ness. There may be an affection of the afferent followed 
by one of the efferent nerves (in the case of reflex 
action), accompanied by no consciousness. 'Sensation' is 
scarcely a proper name for such an affection ; certainly not 
'feeling,' which we always take to imply consciousness. 
(2) Feeling, not individualised as an object, of which the 
expression is a cry of some sort. (3) Presentation of an 
individual object. 2 (2), forming no part of our intelligent 
experience, can only be described by negatives. It carries 
with it no distinction of subject and object; its only dis- 
tinction is as pleasant or painful. How do we know 
anything about it ? Partly by observing actions of our own 
which imply that pleasure or pain is felt, when at the same 
time, from occupation of thought, there is no consciousness 
of pleasure or pain as an object, no reference of it to a thing. 
Partly by observing animals, in which we find the signs of 
feeling, common to them and us, but not the sign of (3), viz. 
language. They cry, but do not afterwards convert their cries 
into signs of things, from which we infer that they do not, 
in the proper sense, perceive things. 

1 [Cf. ib. p. 12, note.] pain already ' intuition'? "Was R Hall, 

2 It may be questioned whether to when, as he supposed, he was uncon- 
animals is to be ascribed only (1), as, scious during preaching of pain caused 
I believe, was the Cartesian view, or by calculus in the kidney, really 
(2) as well. Are they unconscious or without (2) ? Did excitement of 
conscious automata? The cry of cer- thought change (2) into(l)? Or was 
tain animals, I should suppose, showed it that feeling remained, but that he 
them to have (2). Must not (3) be did not attend to it, did not refer it to 
ascribed to them as well ? Does not a himself or a thing ? Probably the 
dog, when it turns to scratch a flea-bite, former. 

localise the pain, and is not the localised 


What Hamilton calls ' sensation proper ' is a form of (3) . 
He distinguishes ' sensitive perception or perception simply ' 
into two kinds (a), 'sensation proper, conversant about 
a subject-object/ and (b) 'perception proper, conversant 
about an object-object.' It is 'that act of consciousness 
whereby we apprehend in our body (a) certain special affec- 
tions, whereof as an animated organism it is contingently 
susceptible (sensation proper), and (b) those general relations 
of extension, under which as a material organism it neces- 
sarily exists.' ! He quite admits (§ 22) that ' sensation proper,' 
thus understood, is mainly 'an act of intelligence.' The 
distinction corresponds to that of ' outer ' and ' inner ' sense, 
according to Kant's doctrine that 'outer' sense is distin- 
guished from 'inner' as that of which space is the form 
from that of which not space, but only time, is the form. It 
also corresponds to the distinction between secondary and 
primary qualities. 

The objections to this usage of terms are, (a) that we 
want the term ' sensation ' for (1) and (2) above ; (b) that 
' sensation ' naturally carries with it an opposition to ' intelli- 
gence,' whereas according to Hamilton's usage it is purely 
an act of intelligence, only qualified as exercised on occasion 
of an ' affection of an animated organism ' ; (c) that it 
conveys the notion that sensitive perception ('perception 
of sensible qualities') is consciously to the subject of it 
apprehension of an affection of his animated organism. 
This of course is not the case : we perceive colour long 
before we are aware that it is an ' affection of our animated 

8. Is, then, the notion of thought given above the true one, 
viz. that it is the function (a) of abstracting attributes found 
in individual things presented to it, and then (b) of taking 
such attributes, fixed by a name, as representative of the 
individual things, which, as thus represented, form classes ? 
If it is, then the developed content of thought may be fitly 
regarded either as a series of groups of attributes, of which 
the more complex contain the less, and expressed by names 
ranging from more full to less full meaning, or, conversely, 
as a series of classes of which the wider, expressed by names 
of less full meaning, contain the narrower, to which names 
of fuller meaning correspond. If this is the true view of 

1 Appendices to Eeid's Works, note D. 

168 LOGIC. 

the content of thought, the doctrine of formal logic is 

In (a), as stated above, there are three fallacies. One 
lies in the antithesis between presentation and representa- 
tion, intuition or perception and conception, as if with the 
former thought had nothing to do. Another, in the notion 
that individual things are ' found ' to have attributes, i.e. 
tha attributes are there and observed apart from the action 
of thought. A third, in the notion that mere abstraction 
of attributes really constitutes any intellectual process what- 

The so-called * immediate intuition ' only has any con- 
tent just so far as it is not merely presentative. Just as, 
when, in view of Locke's doctrine, that that only is ' real ex- 
istence ' which can be known in * actual present sensation,' 
we ask how much of any supposed real object is reducible to 
' actual present sensation,' we find that the object dis- 
appears, so is it when we ask how much of an object of 
intuition remains after abstraction of all that belongs to it 
as representative. ' This book ' is an object of intuition, but 
all qualities in virtue of which I recognise the object as a 
book depend on its relations to objects not now presented in 
intuition at all, of which relations, therefore, the knowledge 
is representative, not presentative. In the absence of these 
nothing remains as merely presented but the ' here ' and 
' there,' ' now ' and ' then,' ' this ' or ' that' ; and can even the 
' this ' and '■ that ' be said to be merely presented ? Does not 
' this ' always indicate a relation of something to, and dis- 
tinction of it from, a subject conscious of itself as not begin- 
ning or ceasing with the presentation of 'this,' through the 
medium of which again the present something is related to, 
and distinguished from, other ' somethings ' P But neither 
the identity of the ego, nor the past somethings to which, 
through common relation to the ego, the present is related, 
can properly be said to be presented. Identity = unity in 
multiplicity. Identity of the ego = its unity in manifold 
experiences : to it the antithesis of presented and represented 
has no application. If it is to be applied at all, we can only 
say that the identity of the self is both presented and 
represented. It is present now, but present as that which 
was equally present in my past represented experience. It 
may be said, indeed, that the other things^ by relation to 


which the ' this ' is determined, are in turn ' presented as 
individual under the condition of distinct existence in space 
or time/ in other words, that ( this here ' is contrasted with 
1 that there,' and ' that, 5 ' that,' and ' that.' But admitting 
that the srspov to any given object is thus a series of 
separately presented individuals, the relation between the 
given ' this ' and them is not ' presented as individual under 
the condition of distinct existence in space or time.' It is 
not in any ' here ' or ' now ' at all, and it is this relation 
which makes the given ' this ' what it is. It thus appears 
that the distinction between intuition and conception, as- 
severally presentative and representative, breaks down. 
Mansel (p. 13) admits that one sort of intuition, viz. ima- 
gination, is both presentative and representative. But the 
possibility of the distinction implies that both the per- 
ceived object and the imagined object are determined by 
relations which are not ' presented.' Merely as" an c act of 
consciousness of which the immediate object is an individual, 
&c.' imagination does not differ from perception, except 
indefinitely (as Hume said) in degree of liveliness. The 
difference is that in the one case the modification of con- 
sciousness, in virtue of certain relations, is referred to a 
permanent cause which it is supposed would operate equally 
upon others, in the other case it is referred to an ' inward * 
cause, to a certain state of my organs which others, under 
like outward conditions, would not share, while at the same 
time its similarity to previous perceptions (in the sense 
defined) is recognised. But neither the relation to an 
' outward ' cause in the one case, nor to an ' inward ' in the 
other, is matter of intuition according to MansePs definition 
of intuition. 

9. Conception is distinguished from intuition as the 
representation of an object under attributes from the 
presentation of an individual object under the condition of 
distinct existence in space or time. But the object ' repre- 
sented under attributes,' according to Mansel, is an individual 
object. There remain, then, according to Mansel, only two 
points in which the conceived object can differ from the 
intuited. The conceived object, though individual, may not 
be 'distinct in space or time,' or the intuited object may be 
presented without attributes. Mansel, however, distinctly 
siivs that in conception attributes must be referred to an 

170 LOGIC. 

object of intuition (it being possible, according to him, to 
have intuition without conception, but not conception with- 
out intuition). It remains that the intuited object differs 
from the conceived in the absence of attributes, while yet 
all the while the content of the conception is supposed to be 
attributes found in, and abstracted from, objects of intuition. 
Thus we are again brought to the same conclusion that an 
unconceived object of intuition (an object of intuition not 
determined by thought) would be a nonentity, and that the 
representation of the process by which knowledge is formed 
as one which begins with such mere intuition and goes on 
to conceptions by abstraction of attributes, is self-contra- 
dictory ; it represents abstraction to take place when as yet 
there is nothing to be abstracted. 

10. In truth attributes mean relations; conception = the 
thought of objects under relations, and under relations every 
object must be thought in order to be an individual object 
at all. The ' thisness ' and ' thatness,' ' here ' and ' now/ of 
the object of intuition are already relations of which the 
intuiting subject is conscious, and of which, as of all other 
relations, he is only conscious because he thinks ; because, 
as a subject equally present to and distinct from successive 
feelings, he holds them together as one. We must observe 
that to be in time and to be conscious of time are different 
and mutually exclusive things. It may be asked, How can 
this be ? When I think of time, is not the thought an act 
of consciousness — an event — which takes place in time ? No 
doubt an act of consciousness is an event in the individual's 
history which is in time, which begins and ends, succeeds 
another, and is over before yet another begins ; but it would 
not be a thought of time but for its determination by a 
subject which holds past and present together, which is no 
more now than it was then or will be to-morrow, and this is 
not in time. The thought of time, like all thought, is 
eternal, but associates itself in man with occurrences in the 
way of feeling which, in virtue of that association, are not 
merely events in time, but are thought of as such. Of two 
successive feelings, one over before the next begins, neither 
can be consciousness of time as a relation between the two. 
Every animal has experiences in time, and animals that see 
have experiences in space in the sense that there are pictures 
on the retina of their eyes of which to us the parts are ex- 


ternal to each other. But it is quite another thing to be 
conscious of relations of space and time. This they are not 
unless, in virtue of other than a feeling consciousness, they 
can hold together (a) successive feelings so as to be conscious 
of them as related in the way of succession, and (b) successive 
acts of vision in which a surface is traversed so as to regard 
them as coexisting and mutually limiting parts of a whole. 
It may very well be (if they do not think, it is so), that for 
no animal does space or time exist, though really and for us 
their experiences are in space and time. 

11. Thought, then, as consciousness of determination by 
relations, is necessary to constitute the object of intuition, 
and if one is to use the antithesis of presentation and repre- 
sentation at all, one must say that thought is representative, 
because neither its object nor the relations by which its 
object is determined are present as feelings. They are 
things which, not being in succession at all, cannot properly 
be said to be either past or present. Here is this table now 
before me. The sensation it excites in me is in time : I 
turn my head and it is gone, 

Like a bubble on a river, 

A moment here, then lost for ever. 

But the relations which make this appearance what it is 
really and in my understanding do not come and go, nor 
does the object, distinct from all others, 'itself and not 
another,' e individual,' which they together constitute. This 
individual object is a thing of the understanding, other than 
the feeling excited in me when I see or touch the table, so 
that I judge it to exist when I am not there to feel it, and, 
when again I have a sight or touch like the first, pronounce 
it to be the same table, whereas a feeling as such can no 
more be the same with another than one moment of time 
the same with the preceding. When I say ' this is the same 
feeling that I had before,' I am thinking the feeling ; the 
feeling is s objectified,' become a thing of the understanding. 
It is not that there are two tables (as Plato dreamt), a ' real ' 
table which I see and feel, an * ideal ' table which I think. 
This individual table which I see and feel (to which I refer 
my sensations) is constituted by relations to the system of 
the universe and to my nervous organism which, like all 
relations, only exist for a thinking consciousness ; relations 

172 LOGIC. 

which would not exist for me any more than time exists for 
a jelly fish which expands and contracts itself in successive 
times, unless I thought ; which do not exist for me in their 
fulness because I think inadequately ; and which, like 
thought, are not in time as feelings are. Plato's mistake 
lay in the confusion of feeling and felt thing, and the con- 
sequent notion that, because feelings were in perpetual flux, 
therefore c sensible things ' were, in opposition to ideal things 
which were eternal. In truth the sensible thing, whether by 
this we mean the sum of the conditions of the given sensa- 
tions or the object to which I refer the sensations, = the 
ideal thing or thing constituted by thought, and is eternal. 
' But,' it will be said, ' the table decays.' True, but decays 
according to eternal laws. That which it was once in certain 
relations it is for ever in those relations. The whole of 
nature does not change, but only the distribution of its parts, 
and that relation to the whole of nature which makes this 
table what it now is determines also its decay, a decay in 
which there is no loss, but only transference. 

12. After all, however, the conviction will remain that 
there is a difference between intuition and conception ; 
between ' ideas which force themselves upon me whether I 
will or no,' and those which I call up at pleasure ; between 
this table as I see it, and my thought of it under its essential 
qualities. These several differences are not to be treated as 
equivalent to each other. There is no doubt a difference (a) 
between intuition and conception, but, according to the upshot 
of Kant's doctrine, 1 it is a difference between thought under 
certain conditions (conditions which arise out of its relation 
to or operation upon sense), and thought exempt from these 
conditions. There is a difference again (b) between imagina- 
tion and perception, as between that which is in a certain 

1 There are not two objects, a con- mination becomes an object, is a datum 

ceived and an intuited ; conception, as of sense. Out of this relation of the 

consciousness of relations, is necessary known object to the datum of sense 

to that simplest individualisation of arises its separateness in space and 

feeling without which there is no ' this ' time, which may thus be said to belong 

or • that.' But (according to Kant) to it as intuited, while it is neutralised 

every known object involves acts both so far as the object is conceived (or, in 

of intuition and conception ; of concep- other words, is properly an object), for, 

Hon, in so far as it is a determination in respect of its determination by other 

of an object through relations to a objects, though 'intuited' as being 

potential universe of objects (a deter- separate from them in time and space, 

mination without which there would it is not really so : all that belongs to 

be no object) ; of intuition, in so far as it, or makes it what it is, comes from 

the matter, which through such deter- them. 


sense arbitrary and that which I cannot help ; and again (c) 
between the real thing and my conception of it, as this at any 
time happens to stand; but (a), (b), and (c) are not to be 
confused with each other. If anyone's conception of an 
object differs from its reality, so must his intuition of it, 
for, as we have seen, all the content of an intuition lies in 
conceived relations. A conception being the thought of an 
object under relations, in intuition the object so thought is 
regarded under special conditions of distinctness in space or 
time. This is the common distinction of all intuition, 
whether ' pure ' or ' empirical,' whether perception or imagi- 
nation. The question is, whether this is the condition of all 
thought, whether I can think an object otherwise than in 
space and time. We have tried to show the mistake of 
supposing that there can be intuition without conception : 
can there be conception without intuition ? Kant held that 
there could be, but that such mere conception or thought 
did not amount to knowledge. Hence, according to him, 
though you could think such objects as a cosmos (the totality 
of conditions), freedom (causa sui), God (the self-conscious 
subject of the physical and moral worlds), and though you 
might be sure that there were such objects (which he held 
to be the case at least with regard to freedom), you could 
not know them, because from the nature of the case they 
were not presentable as intuitions, i.e. as distinct in space or 
time. Kant was quite aware that all relations are conceived, 
not intuited (or, more properly, are conceptions). The 
1 categories,' with him, as those universal relations without 
which there would be no objective world, are * pure concep- 
tions ' to which no object of intuition corresponds, but they 
differ from ' pure ideas,' as being relations, or conceptions of 
relations, which may and do obtain between objects of in- 
tuition ; whereas the * ideas ' of freedom, totality of con- 
ditions, &c. are ideas of that which, from the very nature of 
objects of intuition, can have no application to them. 

13. Mansel l holds (what is quite different from Kant's 
view) that there can be no conception without intuition; 
that, attributes being represented by verbal signs, we may 
and do reason by means of these signs, without at each step 
referring the attribute signified to an object of intuition, 
but that in so doing we reason without distinctly conceiving 

1 Prolegomena Logica, p. 30, ff. 

174 LOGIC. 

what we are reasoning about, and that if at any stage in the 
process, for fear of being misled by mere words, we ask our- 
selves what it is that we mean, we can only answer by an act 
of conception which involves reference of attributes to an 
object of intuition, such reference being necessary to concep- 
tion as such, not merely to true conception. In this doctrine 
there seems to be a confusion between (a) the logical neces- 
sity of referring all attributes to a subject individualised by 
its attributes, of thinking all relations as relations between 
things determined by them, and (b) the necessity, if judgments 
are to represent facts, that they should relate to objects of 
possible perception, should be verifiable by sensible experience. 
The individualised subject in the former case, the subject 
which all conception implies, is not intuited, not necessarily 
presented as separate in space or time. But though it is not 
necessary to conception that its subject should be intuited, it 
may be necessary, if conception is to represent real knowledge, 
that it should relate to experience in the way of sense, to 
objects that can be perceived ; and the perceived object, as 
arising out of data of sense, though not itself such a datum, 
is intuited, i.e. presented as distinct in space and time, though 
at the same time in virtue of the conception (determination 
by relation) the distinctness is denied, suppressed, neutralised. 
It does not follow, however, because a judgment about nature, 
in order to be true, must thus relate to sensible experience, 
that its subject need be intuited or an object of possible per- 

14. It seems to me that the need of intuitionalising 
conceptions depends entirely on the nature of the object to 
which any given conception purports to relate. If it is an 
object of the physical world — this world consisting of phe- 
nomena distinct in space and time — I must be always able, 
if my conception is more than a name, to refer the conceived 
relations either to an object distinguished from all others as 
occupying a certain space, or to an event distinguished from 
all other events as occurring at a certain time. If I talk of 
some kind of reciprocal action between bodies, and my talk 
is to mean anything, I must conceive such action as between 
bodies which I present to the mind's eye in distinct spaces ; 
and further, if my language is to represent not only a con- 
ception but a true conception, I must be able to compass a 
perception of such bodies so acting. (Of this more below.) 


In like manner, if I talk of a particular relation of ante- 
cedence and sequence, to secure meaning I must present to 
myself events in time between which the relation holds, and 
to secure truth of conception I must be able to "perceive 
them. On the other hand, when the conceived relations do 
not purport to be relations between separate things in the 
physical world, e.g. the relation between man and man, 
between subject and object, between motives and will, be- 
tween man and God, or God and the world, there is a 
tendency to intuitionalise the conception arising from the 
fact (a) that it is hard (some say impossible) to think with- 
out expressing thought in language, and (b) that our 
language is primarily appropriate to the physical world 
(indeed to our first impressions of the physical world), and 
that only by a constantly shifting process of metaphor is it 
made to do other duty. This tendency (which is the ultimate 
source of dogma) leads us into paradoxes and contradictions, 
out of which we are apt to find an escape in mysticism. 
The true way of escape is to recognise the tendency itself as 
altogether misleading. Is not the true notion of ' faith,' 
that it is the apprehension of objects which we conceive but 
cannot present in intuition, an apprehension of which the 
proper expression is not language but moral action ? Such 
' faith ' is almost the opposite of what is apt to be reckoned 
so, viz. a facility in presenting intuitions which purport to 
be of that which does not really admit of being so presented 
at all. In moral action, too, not in perception, lies the 
verification of such conception. 'That which for man is 
true, man can verify ■ ; but all verification does not lie in 
the possibility of perception. The existence of the moral 
law and determination by the conception of it cannot be 
verified in this way. There is no perceivable object which, 
as perceivable, demonstrates the existence of the moral law 
because it is impossible otherwise to be accounted for. To 
perception in the strict sense (the same sense in which we 
verify physical theories by it) there may be no difference 
between an act determined by physical causes, an act de- 
termined by expectation of pleasure, and an act determined 
by the conception of a moral law. You cannot, as in 
verifying a theory of physical causation, say, here is an 
action on one side, and here on the other is such a motive 
the only one present in all cases where the act is done 

176 LOGIC. 

therefore this motive must be the cause. You cannot thus 
verify, because the action as ' Handlung ' (' action on its 
inner side ') is not perceivable at all. But every man can 
verify the existence of the moral law by acting according to 
his conception of it ; nay, he even verifies it (shows it to be 
real) by asking whether there is such a law, and why he 
should conform to it ; for its ' real existence ' only purports 
to be an existence for his consciousness, and that it so exists 
he shows by asking such questions about it. And this 
verification of the moral law and of man's capacity for being 
determined by the conception of it (even though the deter- 
mination by it be not such as to outweigh determination by 
sensuous desire) is also verification of a reason which is at 
once author and subject of such law, and which is thus other 
than natural. 

15. So much for the distinction between intuition and 
conception. Now for that between perception and imagina- 
tion. There is undoubtedly a difference between this room 
as I look round and the same as I might represent it to 
myself elsewhere an hour hence. Is the one consciousness 
real, the other unreal ? Is the one the work of ' things 
without us,' the other the ' work of our own mind ' ? If my 
perception of this room were merely a sensation or succession 
of sensations, then its difference from imagination of the 
room, in which there need be no element of sensation at all, 
would be obvious. But there is no perception without an 
intellectual interpretation of sensation. In the supposed 
case, it means that on occasion of a certain sense of colour, 
-a complex object, determined by certain relations, which has 
gradually formed itself in my thoughts, recurs to my con- 
sciousness. In what does this differ from imagination of the 
room? (a) The occasion of the recurrence of the thought 
object to consciousness is different. Imagination is often 
quite as involuntary as perception, but the nervous irritation 
which occasions it is of a different sort, (b) The relations by 
which the object is determined in the two cases, though in 
some respects alike, are in others different. The perceived 
object stands in a certain relation to my body and to other 
things outside my body, in which the imagined object does 
not. These differences between perceived and imagined 
objects, however, are not differences either (a) between ' the 
work of things without us ' and ' the work of mind,' or (b) 


between 'real' and e unreal.' Not (a), because on the otae 
band to the act of imagination as much as to the act of 
perception there is necessary some affection of the animal 
organism, and on the other an intellectual synthesis is 
as necessary to constitute the relations which render the 
perceived object what it is as it is to constitute the imagined 
object what it is. Not (b), because each object has its own 
reality, there being no unreality in the imagined object till 
it is confused with the perceived, till those relations to my 
body and other things outside my body which characterise 
the perceived are wrongly ascribed to the imagined. 1 When 
we ask, Was such and such an appearance real or imaginary 
(in absence of any suspicion of fraud) ? the question is not 
properly whether the appearance has any reality at all (if 
it is ' bare vision,' still it has its own reality), but whether 
the relations of the appearing object (which constitute the 
reality) are such as the subject of the appearance takes them 
to be. 

16. Next as to the difference between the real thing and 
my conception of it as this at any time stands. Our crude 
notion of the antithesis between what is real and what is 
thought gives way before the consideration that all reality 
lies in relations, and that only for a thinking consciousness 
do relations exist. It is apt to be supposed that reality in 
some special sense belongs (a) to feeling, as that which the 
individual cannot help having, (b) to what is material. But 
the supposition (a) in fact means that the feeling is real in 
virtue of its relation to an outward cause, and for a merely 
feeling consciousness there would be no such relation. 
Feelings being successive, there could be no identification of 
one with another (in the judgment * this that I now feel is 
the same object that I felt before '), no reference of feeling 
to an outward cause which does not pass along with it. We 
must always bear in mind that when certain writers speak 
of the c unreality of mere feeling,' they mean feeling as it 
would be for a merely feeling consciousness. Every feeling 
has abundant reality as determined by its actual conditions 
and effects ; but what is meant is that for a subject which 
merely felt there would not be this determination (this 
determination would not be presented as an object). 

17. As to the supposition (6), ask yourself what you mean 

1 Cf. General Introduction to Hume, §§ 188, 189 ; vol. i. pp. 153, ff. 
VOL. II. x 

178 LOGIC. 

by ' matter,' and you find that, whatever the answer (e.g. 
matter is the extended, matter is solid, matter is the un- 
known cause of sensations), it is a statement of some rela- 
tions or other. Thus, if it be true that whatever is real is 
so in virtue of its being material, this still implies that 
reality is constituted by relations, though all but a particular 
sort of relations are arbitrarily excluded. In the doctrine 
of evolution ' matter ' is no longer merely defined by certain 
specific qualities (extension or solidity) that belong to it as 
matter. It becomes what v\rj (in one of its senses) was to 
Aristotle, the germ or possibility (hvvapis) of all things. 
But, according to this view, for the determination of 
matter by certain limited relations is substituted its 
determination by relation to all that is developed out of it. 
Suppose all the life of the universe, including our spiritual 
life, to have been developed out of a primitive matter ; 
the actuality of this must have lain in what has been and is 
being developed out of it. In other words, so far as it was 
really anything at all, it had really a spiritual life. There 
is no alternative between saying that it was really all this 
and saying that it was really nothing. That which is the 
bare possibility of all things can only be in and for itself 
nothing. You cannot say anything of what it is, but only 
of what it is to be. But ' ex nihilo nihil fit. 5 If your primi- 
tive matter were really what it is, merely in and for itself, 
void of qualities, nothing, the evolution of the universe from 
it would be unaccountable. We have therefore to suppose 
that it does not exist merely in and for itself, but for a 
thinking subject for which it is not nothing but all things, 
determined by relation to all that is to be consequent on it 
or come out of it. Thus, though the doctrine of evolution 
is fatal to the old natural theology, so far as. this regarded 
God as a great architect who made the world as a man 
makes a machine, it logically necessitates the existence of 
an eternal thinking subject, in relation to which alone the 
primitive matter is the possibility of what it becomes, and 
each lower phase of life the antecedent condition of a higher. 
Any notion to the contrary arises from the transference to 
matter, in the only proper sense in which the evolutionist 
can regard it (viz. as that which by itself is nothing, though 
the possibility of all things), of the qualities which belong 
to it in the sense in which the mechanical philosopher 


speaks of it, as that which is extended, or resists pressure, 
or conveys force. Transferring these properties to the 
primitive matter, people are able to evade the alternative of 
regarding it either as absolutely nothing, ' ex quo nihil fieri 
potest,' or as being already to an eternal mind what it shall 
become. But if we allow this transference, and so, finding a 
reality for the primitive matter in these qualities, dispense 
with that determination of it by what it shall be, which 
implies its presence to an eternal mind, we are bound to 
show how these properties of matter account for what is 
otherwise unaccountable, life, animal and spiritual, &c. 
which, it must be supposed, not having existed up to a 
certain time, then began to exist in virtue of the properties 
of matter as mere modes of the motion of particles. Accord- 
ing to the true view, which regards ' primitive matter ' as 
indeed nothing by itself, but for an eternal mind determined 
by all that is to follow it, this difficulty does not arise ; for 
according to it, as the ' primitive matter ' is already more 
than what it is in ' and by itself, so every successive actuali- 
sation of it is really what as yet in time it is not. In time 
there was motion before there was organic life, and one sort 
of organic life before another, but really (for the eternal 
mind for which alone it was anything) the motion was more 
than it was in and by itself, was determined by its ts\os, 
had its essence in that which was to follow it. 

18. That * all reality lies in relations ' will more readily 
be admitted than that c only for a thinking consciousness do 
relations exist.' ' Granting,' it may be said (what indeed 
is quite clear), ' that if we did not think as well as feel, if 
there were no thinking subject to hold our feelings together, 
we could not be conscious of relations, it is quite another 
matter to say that there would be no relations.' But at any 
rate one should think that the burden of proof lies with 
those who hold that relations exist otherwise than as we 
know them to exist. For us it is quite certain that only 
through the equal presence to successive feelings of a subject 
other than they, which holds them together, and thus held 
together regards them as its object, are there related things 
or relations at all. It is not that first there are relations and 
then they are conceived. Every relation is constituted by an 
act of conception. 

This is not to be understood as meaning that there is 

N 2 

180 LOGIC. 

' nothing but the soul and its feelings,' or that realities are 
feelings, even feelings as determined by thought. It is 
through feeling as determined by thought that for us there 
comes to be reality, but the reality is not to be identified 
with the process by which we, as thinking animals, arrive 
at it. Even simple facts of feeling (e.g. the fact that a 
certain sweet smell accompanies the sight of a rose) are not 
feelings as felt: more clearly, the conditions of such facts 
are not feelings, even as determined by thought. A c feeling 
determined by thought' would probably mean a feeling 
which but for thought I should not have, e.g. emotion at 
the spectacle of a tragedy. Objective facts are not of 
this sort, not feelings determined by thought, though but 
for the determination of feeling by thought they would not 
exist for our consciousness. 'Is not this to give up the 
doctrine that the reality of the world, as well as our know- 
ledge, is rendered possible by thought ? ' No ; it still 
remains true that * reality and objectivity ' have no meaning 
save as expressing a relation which without thought could 
not be. The world before there was sentient life, was not 
what it is to us as sentient ; the world of conditions of 
feeling is not to intelligence (even our intelligence) what 
it is to us as feeling : but as a world, as real, as objective, 
such a world was or is only what it is to intelligence, to a 
thinking subject, and could not be apart from such a subject. 
19. 1 What true meaning is there in saying that ' sensations 
are in flux'? Does not psychology teach that each sensa- 
tion is 'registered,' remains in effect as a modification, 
however slight, of the ' pyschoplasm,' which qualifies every 
succeeding sensation ? The answer is, that sensation con- 
tributes nothing to the ' cosmos of our experience,' is not a 
possible subject of relations, except so far as attended to. 
As attended to, it is a passing event, related to and deter- 
mined by, forming one series of change with, former and after 
events, and this in virtue of presence to and distinction 
from a thinking subject. The permanence of (a) the effect 
of the sensation is not a permanence of what the sensation 
was to me, as (b) an event in the way of feeling. What 
memory retains or recalls is (b), not (a), (a) may be 
permanent, and there may yet be no memory. Unless 

1 [This section is out of place here; note to the latter part of section 24, 
it would come more appropriately as a below.] 


sensations were in flux in sense (b), we could not be conscious 
of them as changes, any more than we could be so unless 
there were something beside them not in flux. The qualifica- 
tion of a sequent sensation a? by a previous one y is not a 
consciousness of them as events or of a? as a change from 
y. Hence permanence in the psychical effect of sensation 
does not in the least facilitate the reduction of intelligence 
to sensation. If sensation, as it is for consciousness, or as 
a hvva/juis of knowledge, were permanent, it could not be 
such a hvvafus. It must pass, yet be retained as having 
passed, in order to become a factor in any of the most 
elementary relations which are conditions of knowledge. 

20. The objections which suggest themselves to the 
doctrine that relations are constituted by thought do not 
apply to the doctrine itself (which, once understood, is 
irrefutable), but to its supposed implications, (a) What, 
according to it, becomes of ' external matter,' which all the 
exact sciences suppose ? The answer is, that it is unaffected 
by the doctrine, except that ' externality ' has to be under- 
stood as of matter to matter, not of matter to thought, 
6 matter ' and ' externality ' alike meaning certain relations 
which thought constitutes, (b) Is there then nothing other 
than thought ? (c) Is the universe the creation of my own 
mind? How can that be when I only began to think 
twenty-five years ago ? We answer to (b) ; undoubtedly there 
is something other than thought. Feeling is so ; the whole 
system of nature, on which feeling depends, is so; its 
otherness from thought makes it what it is, but this is the 
same as saying that relation to thought makes it what it is, 
that but for thought it would not be. Conversely, c other- 
ness ' from nature makes thought what it is. The very idea 
of thought implies a sTspov, for thought = self-consciousness, 
or consciousness of the distinction between subject and 
object, and thought cannot be conscious of itself except in 
distinction from an object. The mistake lies in a confusion 
between the relation of object to subject, and that supposed 
externality of matter to mind, which is really a transfer of 
the true externality of space to space to the relation between 
subject and object. 1 Subject and object, thought and its 

1 It is important not to confuse organism. It is a common delusion 
the relation of subject and object with that one sort of phenomena are 'sub- 
the relation of matter to the psychical jective,' another ' objective.' In truth, 

182 LOGIC. 

srspov, are correlative or complementary factors in the whole 
of self-consciousness, or (which is the same) together con- 
stitute the reality of the world. Each is what it is only in 
relation to the other, but there is this difference, that whereas 
it is true to say that only for the subject or for thought 
is the object, or the srspov, what it is, it is not true to say 
that only for the object or for the srspov is the subject or 
thought what it is (just because the 'for 9 implies relation to 
consciousness, and the srspov is that in the whole formed by 
self-consciousness which is not conscious). 

21. (c) The point of this objection lies in the sense of 
exclusion attached to c my own/ and in the supposition that 
'my mind' began to exist when I (my animal organism) 
was born, or perhaps not till somewhat later. 1 What is 
meant on the other side is that the world, with the whole 
process of development in time (including the development 
out of lower forms of that animal organism which is the 
hvvapLLs of thought), exists eternally for an eternal mind, 
or as a factor in an eternal self-consciousness; that this 
eternal mind uses the animal organism in man as its vehicle 
so as to constitute a being self-conscious, yet limited by 
conditions of the organism in respect of the srspov (object- 
matter) which its self-consciousness can comprehend; that 
thus the human mind, qua mind, has not a beginning in 
time at all, does not exist in time at all. That which exists 
in time is the organism fitted to be a vehicle for self- 
conscious thought, and of this the beginning is not properly 
to be fixed at birth or at ' conception in the womb ' ; it has 
had a history of which we seek the beginning in vain. But 
this history itself, — the connected series of events, determined 
by the constant system of nature, which forms it, — as it 
exists only in relation to an eternal and self-conscious mind, 
so is knowable by us only because this mind constitutes the 
6 me ' in each of us ; only because it so uses the animal 
organism of man as to form a being formally self-conscious, 
and thus capable of knowledge, able to conceive a world of 
which each element is determined by relation to the whole, 
though but slowly advancing to the articulation of this 

' mental phenomena' are just as objec- and object are correlative factors of 

tive as any, phsenomena of matter just everything as known, 
as subjective as any. If mind and ' See General Introduction to Hume, 

matter = two orders of phenomena, they § 129 to § 152, especially §§ 133, 134 ; 

do not = subject and object, for subject vol. i. pp. 113-115. 


22. But it may be said, c Is not all this a mere guess, 
due to man's inability to view the world save through the 
coloured glasses of his own subjectivity?' On any other 
supposition the world which we know, the world which can 
be shown to be determined by relations which thought 
constitutes, must (a) be one which begins and ends with the 
birth and death of the individual man; and (b), if we still 
cannot help supposing that there is a real world, a world of 
* things in themselves,' out of relation to what we know, its 
existence will just suffice to render what we know unreal, 
but in every other respect will be for us nothing at all. We 
vainly try to evade (a) by saying that, though the intelligence 
of each individual, through which there is a present world 
for him, begins with his birth and ends with his death, yet 
from this he rightly infers the existence of a past world. 
The thing inferred is not, any more than the conditions of 
understanding which determine the inference, outside the 
thinking consciousness. That very relation in the way of 
time which ' past ' expresses, like all such relations, arises 
from the presence of the ego to feelings, and cannot other- 
wise be accounted for. If ' my own mind ' began with birth 
and ended with death, the world of which alone I know 
anything — the inferred past no less than the present — must 
do so likewise. 

23. As regards the supposition (b) of a world of ' things 
in themselves,' it may be said, ' Why may there not be such 
a world? How do I know that the categories of my 
knowledge are forms of the real world ? Why may there 
not be a world undreamt of to which they do not apply ? ' 
A negative to such a suggestion cannot be proved. What 
has to be done is (a) to prevent misapprehension by pointing 
out that the categories (cause and effect, &c.) only purport 
to be laws or forms of an objective world as existing for a 
mind, not to be conditions of mind itself. An objective 
world, it is held, can only exist in virtue of these supreme 
determining relations, and they again only exist for a mind 
or thought, but to thought itself they do not apply. You 
cannot say properly that thought is a cause or effect, a 
substance or an attribute. If, then, you like to speak of 
thought as a ' spiritual world,' and say that the categories 
which regulate the knowledge of nature (in virtue of which 
every consequence has a uniform antecedent) do not apply 

184 LOGIC. 

to the ' spiritual world/ well and good ; but you must not 
then go on to mix up this spiritual world with the natural, 
and talk of 'the supernatural projecting itself into the 
natural/ The ' supernatural ' with most people is a hybrid, 
neither natural nor spiritual, (b) It must be pointed out 
that the supposition of there being another world not deter- 
mined, as that which we know is, by thought, is itself the work 
of self-consciousness, still distinguishing an object from itself, 
but trying to detach from the object all that determination 
which really belongs to it in virtue of its relation to the self. 

24. If it be true, then, that without relations there is 
no reality, and that only for a thinking consciousness do 
relations exist, what becomes of the difference between ' real 
things ' and our conceptions of them ? ' Everybody ' (it will 
be said) 6 knows that my conception of a flower, however 
correct, is not the real flower.' Quite so ; but why not? (a) 
, The conception is my own making, but I defy you to make 
a flower.' (b) * I can see, smell, and touch the flower, but 
not my conception of it.' (c) ' The flower is an individual 
thing ; my conception of it is not, but only a representation 
of such through its attributes.' 

Take (b) first. The sight, &c. of the flower means that 
certain sensations have been referred to an identical thing, 
a thing thought of as one throughout successive sensations, 
which thus becomes the subject of sensible properties ; and 
that on the recurrence of any such sensation, it is inter- 
preted as a sign of this thing, so that the mere sensation is 
immediately superseded by the judgment, ' I smell a flower.' 
Now all this is the work of the understanding, apart from 
which there might be this or that smell, sight, &c. but not 
the flower which I see, &c. because there would be no unity of 
successive sensations in an identical thing. The real flower, 
then, means certain relations of sensations, which are not 
themselves sensations, relations which only exist as con- 
ceived, in virtue of an intellectual synthesis. The difference 
of it from my conception of it is, that in the latter certain 
relations are detached (a) from sensations which in reality 
they determine, and (b) from a multitude of other relations 
with which they are really connected. The extent of the 
latter detachment depends on the more or less scientific 
character of the conception. Thus our thought of natural 
reality is always in the rear of reality itself for two reasons. 


(1) Sensation is always in flux. 1 A feeling occurs, is deter- 
mined by conceived relations, and as the result of that de- 
termination is, for consciousness, a real thing : the next 
moment it has gone, and it seems that for thought the 
relations alone remain. This implies that sensation is a 
necessary complement of natural reality. ' Then,' it may be 
said, c there is no natural reality before actual sentience.' 
The relation to sensation yet to come is involved in the 
reality of nature as nature was before there were as yet 
sensations. In like manner the relation to sensations of 
yesterday is involved in the reality of today. That sensa- 
tions are past makes no difference in reality, as this is for 
that consciousness for which alone there is reality at all. 
The conceived fact, the reality, that such a sensation occurs 
under such conditions, is unaffected by the circumstance 
that the sensation is not now occurring. (2) Being ourselves 
beings of slowly emerging intelligence, we can never com- 
prehend the relations of the natural world in their fulness, 
and if at this moment we could know all that is, tomorrow 
it would have changed — changed according to intelligible 
laws, but still changed. 

25. 2 The conception of the flower may be opposed (a) to 
actual events in the way of sensation (a certain sight or 
smell, &c. as at any time occurring), as determined by rela- 
tions ; (b) to possibilities of those events. From (b) a true 
and adequate conception does not differ at all. As to (a), 
the mere event of sense is not reality, nor for a subject that 
merely felt the sensation would there be such a thing as 
reality. For a subject perfectly intelligent the difference 
between (a) and (b) would not exist. Reality would be the 
fact that a sensation shall occur or has occurred, just as 
much as that it is now occurring, because such a subject 
would not be a subject of the sensation. But we not only 
have the consciousness for which there exist the relations 
which constitute natural reality (intelligent consciousness) ; 
we have also the sensitive consciousness which is a factor in 
those relations. We must experience the sensation before 
we know the fact that it occurs under certain conditions, 
and constant repetition of sensation is necessary in order to 

1 [Cf. section 19, above.] 

2 [This section seems to be a restatement in a fuller form of the latter half 
of the preceding section.] 

186 LOGIC. 

the completion of our knowledge of the conditions under 
which it occurs. But supposing that knowledge once 
attained, the full reality would exist for us, as known fact, 
without repetition of sensation. It seems as if reality were 
perpetually vanishing, but on consideration we find that 
what has vanished was not the reality. "The form remains, 
the function never dies.' Sensation vanishes, but not the 
fact that it has occurred under certain conditions and leaves 
certain effects, and this is its reality, but a reality only 
possible for a synthetic intelligence. 

26. As to (a), 1 a conception is not ( my own making ' in 
the same sense in which an artificial flower is. It is the 
result of past experience, and though, as just shown, but for 
the activity of thought in this experience it would not be an 
experience of things at all, still this activity is not one 
which it rests with this or that man to exercise or not, as it 
seems to rest with me whether I shall now walk out of this 
room or not. "No doubt I can arbitrarily combine objects in 
conception which cannot be combined in reality (I can con- 
ceive centaurs, &c.) : such conceptions, however, are not in 
question. What is in question is the relation between the 
conception of a real thing and the real thing itself. The 
former is not my own making, but, being made, can be 
retained, as the sensation cannot. The reality of objects 
which we know, or are in process of coming to know, is sense 
determined by thought relations. As explained, sensation 
is that which we cannot retain, and of which we cannot 
by thinking command the recurrence, while the thought 
relations are in a sense our own. What the thinking subject 
has contributed to reality, it retains from reality : in this 
sense the conception of the flower is my own making, while 
the sensation which must supervene upon this conception in 
order to constitute a real flower is an event which I cannot 
command, and which, having occurred, becomes part of a 
past which I cannot reproduce. Adopting the distinction 
between conception and sensation, as one between what I 
can make and what I cannot, we must say that, though what 
I can make does not amount to the real flower, no more does 
what I cannot make. Only if continuance or reproduction 
of feeling were necessary to the reality of the represented 
object, would the opposition lie between the real object as 

1 [Above, section 24.] 


that which I cannot make and the thought of it as that 
which I can. 

27. The same considerations help us to deal with (c). 
What is meant by the individuality of the real thing ? Does 
it mean (1) singleness, distinctness in time, of the sensitive 
act in which it is presented ? Or (2) its identity, that in 
virtue of which it is itself and not another? Or (3) the 
complex or combined result of the relations which determine 
it or make it what it is ? Or (4) that special result of such 
relations which consists in organic life ? As to (1), we have 
admitted that the sensitive act is other than any such rela- 
tion as thought constitutes, and that it is necessary to the 
reality of the natural thing. It is an event in time, and, as 
such, the absolute srspov to self-contained thought. But 
then we must remember that, as soon as we speak of a 
moment of time, or of an event as occurring in such a 
moment, we have intellectually determined it by a relation 
only possible for a subject that holds successive moments 
and events together in one. Thus in the first sense of 
individuality (distinctness of a sensation in time), we may 
hold that it is necessary to reality, and that it implies the 
presence to thought of something other than thought, but 
which yet derives its determination from thought ; and of 
this abstraction is made in conception. In sense (2), indi- 
viduality is a pure datum of thought. A feeling indeed 
may have identity, but only as converted into a felt thing, 
as retained by thought from and after the act of sense, and 
then contrasted with other objects so retained. Every con- 
ceived object, as conceived, has identity; no supervention 
of sensation upon the conception is necessary to constitute 
the identity of such an object, or makes any difference to it ; 
and only a conceived object can have it, because without 
conception there would not be the synthesis of differences 
(of one thing from all other things) which constitutes it. If 
this is true of individuality in sense (2), it is, if possible, 
more obviously true of it in sense (3), which indeed is merely 
the actualisation of (2). The individuality which = bare 
identity, the distinction of a thing always the same with 
itself from all other things, is particularised, or made a 
definite sort of identity, in virtue of the specific character of 
its distinction from other things which is at the same time 
its relation to them. As only through the presence of the 

188 LOGIC. 

thinking subject to, and the action of it upon, a sensation 
does this become a separate thing, so only through the same 
presence and action does qualification by relation gather 
upon the thing. It is not in dispute that every sensation 
really has a definite individuality or character, more definite, 
because depending on more complex conditions, than any of 
us knows. What is maintained is that merely as a sensa- 
tion, or apart from relation to a thinking subject, it would 
not be so. Thus, though a feeling as it occurs to the think- 
ing man is individual, it is so not in virtue of what it merely 
is as a feeling, but in virtue of what it suggests. In other 
words, the individuality belongs to the thing of thought, to 
the conceived object, which does not become any more 
individual because a sensation occurs, which I refer to it. 
The conceived object which I call ' the pear-tree in my 
garden ' is no less individual as thought of than when I see 
it. The act of seeing it no doubt is individual in time, which 
the thing of thought is not, but to individuality in senses 
(2) and (3) the occurrence of sensation makes no difference. 
28. It is a mistake, then, to oppose the real thing to the 
conception of it as individual to what is not individual but 
' abstract universal.' The conceived object, the thing indi- 
vidualised by relations which does not come and go with sensa- 
tion, is the only thing. There is no real thing other than it. 
It is not that there is a conceived thing which is unreal, and 
a different perceived thing which is real. Qua thing there is 
no difference between the conceived and the perceived thing ; 
and in perception it is thought, or an act of conception, in 
virtue of which alone there is a thing perceived. But in per- 
ception there is, over and above the relations which deter- 
mine the thing as conceived, a relation to the sensitive 
organism of the perceiving (which is also the conceiving) 
subject. The possibility of such a relation is the test of 
natural reality; i.e. the only real objects and relations in 
nature are such as are either more nearly or more remotely 
related to sense. A conceived past condition of the earth, 
such as the geologist describes, is real because it is the only 
possible explanation of some part of our present sensitive ex- 
perience. But the possibility of such explanation presupposes 
the identity of nature, — that nature is one through all time, 
so that all phenomena are changed appearances of what 
remains the same. Thus in granting that perception is 


the sole verification of conception, we must (a) restrict this to 
conception about nature ; (b) note that perception itself is not 
mere sensation or feeling but relation of something to sense 
only possible for a conceiving mind; (c) that a conceived 
relation verified by perception is seldom itself a relation to 
sense given in perception; and (d) that the possibility of 
verification depends on the pure conception of the unity of 
nature, on the conception of what can never be perceived. 

It is not that first there is conception, that then sen- 
sation supervenes, and reality results. Natural reality 
involves relation to sense, and the occurrence of sensation 
ex parte nostra is the test whether our conception of reality, 
which as a conception of reality implies the belief that a 
certain sensation will occur under conditions, corresponds 
to reality or no. It is not that any particular reality first 
comes into being on the occurrence of my sensation. 1 

The real thing, then, is individual because universal : 
i.e. its individuality lies in its relation to all other things, 
which is a one in all, the common element in all, a universal ; 
it lies in this relation, this mere difference from all other things, 
as particularised. Thus we may not say either that the real 
thing is individual, not universal (for its individuality is 
a universal particularised), or that its individuality dis- 
tinguishes it from such a work of thought as conception (for 
its individuality is the work of thought) . If, however, we hold 
that the thing as real differs from the thing as merely con- 
ceived in virtue of actual or possible relation to sense (which 
may be true of natural 2 reality), then, since such relation 
implies occurrence in time, we may truly say that individu- 
ality in time is an incident of realisation ; that, every ' thing * 
being a conceived thing, the possibility of presentation in a 
distinct time is necessary to its being also a real thing. 

29. Neither the thing as we at any time conceive it, nor 

1 Can relation to sense, as a fact or (inward) act, not one which we find real 
reality, exist for a consciousness not through occurrence of sensations. The 
sensitive? If not, how do facts of nature conception of an absolute good, of a 
exist for God ? ' categorical ' law, of freedom (moral 

2 ' Natural ' reality only ; to the moral autonomy), is equally real, so long as it 
object this distinction between the real determines the will, though no object 
and the merely conceived does not apply. can be found corresponding to it, as 
That which must supervene upon the from the nature of the case there cannot 
mere conception of it in order to it3 be, for what is found given in sensitive 
becoming real, is determination of the experience can only be an event con- 
will. In other words, a moral concep- ditioned by another event, 
tion is one to which we give reality in 

190 LOGIC. 

the thing as we feel it, is the thing in the fulness of its 
reality. I have a conception of a flower, and upon the oc- 
currence of a sensation, which I interpret by means of this 
conception, I judge 'there is a real flower'; but the flower 
is really much more than the relations which I had previously 
conceived plus the present relation to sense. But this ' more ' 
still lies in relations which can only exist for a conceiving 
mind, and which my mind is in process of appropriating. 
The great mistake lies in regarding a conception as a fixed 
quantity, a ■ bundle of attributes.' In truth a conception, 
as the thought of an object under relations, is from its very 
nature in constant expansion. Hence the impossibility of 
really defining a conceived object, unless the relations which 
determine it (like those of space) from their primariness ad- 
mit of being isolated. The ordinary definition of an object 
is available only for rhetorical purposes, as expressing what 
for the time certain disputants, or those to whom a man is 
speaking, agree to understand by a name. 

30. 1 Would a perfectly adequate conception of the condi- 
tions of a phenomenon (an event in the way of sensation) 
differ from the reality of such phenomenon? (1) The conception 
of conditions is equivalent to the conditions as conceived, since 
the conditions are only possible as constituted by a synthetic 
intelligence, and our conception, so far as adequate, is a 
repetition of the act of such intelligence. If an ' adequate 
conception, &c.' then, is to differ from the reality, it must 
be because something besides the conditions of an event in 
the way of sense is necessary to complete its reality, viz. 
actual sensation. Is this so? Is not the notion that an 
event in the way of sensation is something over and above 
its conditions, a mistake of ours arising 1 from the fact that 
we feel before we know what the reality of the feeling is, and 
hence continue to fancy that the feeling really is something 
apart from its conditions? For the knowing consciousness, 
even in us — that consciousness for which alone there is reality 
at all — the fact that a certain sensation is experienced under 
certain conditions, once learnt, is independent of actual ex- 
perience of the sensation. For it the conditions are the reality 
of the sensation. Eepetition of sensation is only needed from 

1 [This section treats over again, in essentially the same results, the ques- 
a somewhat different form, but with tions discussed in §§ 25 and 26.] 


the infirmity of our minds, in order to fuller knowledge of 
the conditions. To say, then, that the fully conceived con- 
ditions of a phenomenon still fall short of its reality, — that 
sensation must supervene upon them in order to constitute 
reality, — is a mistake. For the only sort of consciousness 
for which there is reality the conceived conditions are the 
reality. ' The conditions, however, are, or include, relation to 
feeling.' True ; but relations to feeling are not feelings, 
and the consciousness for which the full reality of such 
relations exists must be analogous to the consciousness in 
virtue of which we know, not to that in virtue of which we 

(2). As regard our conceptions, they are confessedly 
always inadequate. Does perception give us reality in a 
sense in which conception cannot ? There is no perception 
without conception, but in order to perception sensation must 
supervene upon conception. Is such supervention necessary 
in order that the conceived (as conceived by us) may become, 
or represent, the real ? Subtract from the perceived all that 
which is conceived, as distinct from feeling, and nothing 
remains to be real; but, conversely, subtract from the per- 
ceived that in virtue of which you distinguish it from the 
conceived, would what remains be real ? The given con- 
ception, the thought of an object under certain relations, is 
affected in respect of reality by perception just so far as per- 
ception further determines it — leads us to know the object 
under relations not known before. The perceived object is 
the same as the conceived, but in perception it may become 
more real, just so far as on occasion of a feeling there is 
further thinking. Conception, however, as it remains over 
from perception in which one has learnt something, is not 
related to reality otherwise than the perception. I perceive 
something and then shut my eyes : perception over, con- 
ception remains, and the conceived fact is just as real as the 
perceived. It only ceases to be so, in so far as conceived 
determinations are gradually dropped from it. 

I cannot make the thought of the real object any more 
than the real object. I can in some sort retain the thought 
of the real object, — i.e. my inadequate conceptions of the 
conditions which constitute it, — but the only reason for 
saying that in so doing I do not retain the real object in any 
sense in which I ever perceived it, is that the sensation in 

192 LOGIC. 

which it was supposed to be given vanishes. But though 
the sensation vanishes, the conceived fact of its occurrence, 
which is alone the reality, remains. 

31. What has been so far said has arisen out of the 
proof that the opposition between intuition and thought, 
as between presentative and representative, is fallacious. 
Incidentally it has further shown the fallacy of supposing 
that it is the function of thought, having found things 
with attributes given independently of it, to abstract these. 
Thought, as the faculty of synthesis — holding sensations 
together as a sensible thing, and one thing with another 
as mutually qualified — constitutes the attributes. The process 
of abstraction, as ordinarily described (as beginning with 
complex attributes and leaving out attributes till the notion 
is reached which has the minimum of determination), if it 
really took place, would consist in moving backwards. It 
would be a donkey-race. The man who had gone least way 
in it would have the advantage, in respect of fulness 
and definiteness of thinking, of the man who had gone 
furthest. The false doctrine of abstraction, as we find it in 
Greek philosophy, and as it has descended to us through the 
Scholastic logic, has its root in the conversion of the true 
antithesis between mere feeling and the work of thought into 
the false one between the sensible thing (feeling as deter- 
mined by its conditions) and the work of thought. It being 
true that the process to knowledge is a process from mere 
sensation to the relations which determine it, and which 
exist for thought (or are votjto), Plato through the above 
confusion came to regard it as a process away from sensible 
things — away from those conditions of sensation (the ' ordo 
ad universum ' in distinction from the ' ordo ad nos ') at 
which it is the true object of knowledge to arrive. A pro- 
cess, however, (a) to what ? and (b) by what method ? All 
sensible things — the whole order of phenomena — being ex- 
cluded, nothing remained to be the object or end of the 
process but that of which you cannot say anything in par- 
ticular, pure being. The intellectual process ends where it 
began, with that of which you can only say ' it is.' If there 
is any difference, it is only that between ' there it is ' and the 
mere ' it is.' The method corresponded to that of the Socratic 
search for definitions. The object of this as employed by 
Socrates was to obtain precision in the use of general names 


by considering cases in which they were applied and leaving 
out those features of each case to which the name was not 
relevant. For this (and incidentally for moral self-examina- 
tion) it has its use, but not so when treated as a process by 
which thought, taking its departure from the half reality of 
the sensible thing, reaches true reality. Thought is thus 
reduced to the office of analysing the contents of general 
names. When, with the revival of interest in physical 
knowledge, the notion of reality as consisting in abstractions 
was given up, the notion of thought corresponding to the old 
notion of reality was retained. Hence the antithesis, which 
has governed modern English philosophy, between the work of 
thought and reality. If the function of thought is abstraction , 
the highest idea (as that on which the function of thought has 
been most exercised) will have least reality : in short, the 
more we think, the less we shall know. 

32. If we say, in opposition to such a doctrine, that thought 
is a process from the more abstract to the more concrete, — 
that we know things first under a minimum of qualification 
and afterwards under more, — we seem to contradict the fact 
that knowledge begins with experience of real objects, 
which, as real, are qualified with infinite complexity. ' Can 
you deny (it will be said) that it so begins with experience ; 
or that objects of experience are thus real in the most 
concrete sense?' We answer; it does so begin, and the 
objects are thus real, but only in themselves ; for the subject 
learning to know they are so only potentially, not actually. 
For him the beginning of knowledge is merely, 'there is 
something,' in other words, his first idea is of e mere being ' ; 
this ' something ' gradually becomes further qualified, as, in 
virtue of that relation of the ego to passing feeling which 
renders it ' something,' it is held in relation to other experi- 
ence. Thus ' concrete ' objects are gradually constituted by 
a process which is conjointly one of synthesis and analysis. 
It is not that there is first analysis and then synthesis, or 
vice versa, but that in and with the putting together of 
experiences, the world before us, which is for us to begin 
with confusedly everything and definitely nothing, is resolved 
into distinctness ; or, conversely, that as resolved into dis- 
tinctness, it assumes definite features which can be combined. 
Every sensation attended to implies a detachment of it from 

kthe flux of successive feelings, and so far an analysis by which 

194 LOGIC. 

it and they are alike to a certain extent determined, and also 
a synthesis of it with them. 

33. All speculative thinking and knowing, even of the 
scientific kind, is a continuation of this conjoint analysis and 
synthesis by which, for the first time, knowable things are 
constituted for us. So far as there is a difference between 
two stages of thought and knowledge, it is between the 
later, that in which we know that we are knowing and 
think that we are thinking, and the earlier, that in which 
we do not. This difference is sometimes put in a misleading 
way as that between conscious and unconscious thought. 
(In another sense still, nature is sometimes said to be 
unconscious thought, which, to be true, must be taken to 
mean that natural things only exist for thought, or in 
relation to a self-conscious subject, but do not themselves 
think.) All thought must be conscious (which does not 
mean £ accompanied by any sensitive emotion '), but need 
not imply reflection on itself as thought. In ordinary know- 
ledge there is no such reflection. Hence ordinary men are 
quite unaware of any activity of thought having contributed 
to constitute the things of which they have experience. 
When it comes to scientific inquiry they know they are 
thinking, but, from the preconceived idea that thought has 
contributed nothing to the constitution of the things thought 
about, they give a wrong account to themselves of what their 
scientific thought consists in, and regard it as mere methodic 
reception (suppose that the mind in scientific thought, as in 
ordinary experience, is merely receptive, but more methodi- 
cally receptive). For us the preliminary or * unreflective ' 
stage of knowledge is indefinitely abridged by language. 



(Mill, Book I. Chapter V. 1 ) 

34. Logic, according to Mill, is the science of ' proof,' 
of ' evidence,' of * inference,' i.e. of the process by which we 
arrive at such true ' beliefs • as are not intuitive. 

The account of the distinction between two kinds of 
truth (In trod. § 4) is, ' Truths are known to us in two 
ways ; some are known directly, and of themselves ; some 
through the medium of other truths. The former are the 
subject of Intuition, or Consciousness ; the latter of Infer- 
ence.' Logic only deals with the latter kind, though the 
question what truths are of the latter kind, what of the 
former, is not for the logician to decide (ib.). Truths of 
the latter kind, however, form ' far the greatest portion of 
knowledge,' so that ' nearly the whole of science is amen- 
able to the authority of logic ' (ib. § 5) . Directly afterwards 
we are told that ' the field of logic is coextensive with the 
field of knowledge.' It is different, however, from knowledge, 
since its business is not to arrive at inferred truths, but 
to determine whether they have been arrived at. (It is not 
the process, but the theory of the process.) 

A truth, whether of intuition or inference, is a proposi- 
tion (I. i. § 2), and a proposition expresses belief in a 
matter of fact consisting in a relation between two pheno- 
mena (I. v. § 1). 'Matter of fact,' ' object of belief,' 
'import of a proposition' — even 'proposition' alone — are 
interchangeable with Mill. ' Logic is not the science of 
belief, but of proof (p. 8). 'Nothing,' however, 'but a 
proposition can be an object of belief, or therefore of proof 
(I. iii. § 1). Everything proved, then, must be a belief, but 
every belief is not proved or provable. 

1 [The references are to the 4th edition (1856).] 

o 2 

196 LOGIC : MILL. 

The ' relation between two phenomena ' is also spoken 
of as a relation between two 'nameable things ' (I. i. § 2). 
Accordingly, since before propositions can be proven there 
must be propositions, and before propositions names and 
nameable things, the inquiry into the nature of proof must 
be preceded by an inquiry into (which with Mill means a 
classification of) (a) the meanings of names, (b) the relations 
between nameable things expressed in propositions. 

The order which Mill adopts — beginning with names, 
going on to nameable things, and finally to the import of 
propositions — is essentially misleading. As he himself 
admits, the minimum of 'belief or ' possible truth' is a 
proposition. Nothing less than what can be stated in a 
proposition is a matter of fact at all. Except, then, as a 
constituent in a matter of fact, a ' nameable thing ' has no 
intelligible character. It is for knowledge nothing. Nor 
can ' names ' be classified without reference to that which 
they signify— without reference to ' things,' which are what 
they are in virtue of relations which only propositions can 
state, in other words, which are elements in ' matters of fact ' 
apart from which they are nothing at all. 

35. Thus the right order would have been to begin with 
the import of propositions, analysis of which should have 
supplied a doctrine of categories — a doctrine which Mill 
attempts to supply in the chapter on 'things denoted by 
names ' — and then, having thus answered the question as to 
nameable things in the only way in which it can be answered, 
to classify names according to the function which they fulfil 
in different sorts of proposition. In fact, in Chapters II, 
III, and Y, Mill is dealing with one and the same question, 
but answering it in different and more or less inconsistent 
ways. This is the question, What are those primary re- 
lations without which there would be no world of connected 
matters of fact to be known at all, and of which all other rela- 
tions are determinations, — which form the universal element 
that is particularised in all knowledge ? Such relations are 
conceptions. It is a mistake to speak of them as relations 
other than conceptions, of which we form conceptions. They 
are constituted by the act of conceiving, by the presence 
of the ego to the ' manifold of sense.' They are not the gra- 
dual result of experience, for they are the presuppositions 
of experience. They are implied in the whole process by 


which the human race has come to knowledge, and by which 
the individual now comes to knowledge, but, according to 
the difference already noticed l between knowing and knowing 
that we know, we do not know them during the process 
which they alone render possible. The true account of 
them can only be got by an analysis of knowledge, — by 
e tracing its genesis,' if you like, if by this you mean a 
process from simpler to more complex conceptions, not if you 
mean a process in time by which the human race has 
gradually attained them (such a process as it is the fashion 
to call a 'natural history of knowledge '). What people are 
really tracing when they imagine themselves to be tracing a 
process by which formal conceptions have originated, is the 
process by which they have been gradually disentangled and 
ha ye found abstract expression. Of the latter process a 
history is possible, but it presupposes both {a) a knowledge 
of what these conceptions are, that can only have been 
attained by an analysis of knowledge, well or ill performed ; 
and (b) the immanence of these ideas in the human mind 
during the process by which they have found abstract 
expression. To seek to get at their real significance by 
etymology, by ascertaining the exact nature of the sensuous 
clothing in which they were first invested, is childish. 

36. What has Mill to say about them in the chapter on 
the import of propositions P A proposition expresses ' belief 
having reference to things,' not (as he holds that earlier 
writers, adopting Locke's definition of knowledge, have 
supposed) ' to our ideas of things.' It is quite true that the 
belief expressed by a proposition does not relate to a ' mere 
idea,' in the sense of something which in the act of belief I 
regard as merely belonging to my own mind, as distinguished 
from 'things.' But Locke, in his definition of knowledge, 
did not mean ' idea ' to carry this sense. * Idea ' with him 
= ' the immediate object of the mind in thinking,' and 
those ' ideas ' which in his language are ectypes, not arche- 
types, viz. simple ideas and complex ideas of substances, 2 
correspond very much to Mill's ' phenomena ' as = states 
of consciousness referred to outward things causing them, 
which in their 'real essence' are unknown to us. But 
Locke, in regard to ' ideas of substances ' (e.g. gold), finds 
himself in presence of a difficulty which Mill, by an easy 

1 [Above, sec. 3.3.] 2 See General Introduction to Hume, § 116; vol, i. p. 95, 

198 LOGIC : MILL. 

way of talking of ' phenomena,' evades but does not meet. 
They purport to relate to ' external things/ things outside 
our minds, and of such things we know nothing but the 
ideas they produce in us (the phenomena they exhibit) at 
the time of their production. There is an outward some- 
thing, of which in itself we know nothing, which produces 
in us the sensible effects — yellowness, hardness, &c. — which 
make up our complex idea of the substance gold ; but this 
only entitles us to such judgments as * something is now 
producing in me this sensation of yellowness,' or ' something 
which produces the sense of yellowness also at the same 
time produces that of hardness, &c.' When I come to such 
a judgment as ' gold is yellow ' — a statement not purporting 
to describe a present sensation — I have gone beyond any- 
thing that I am entitled to assert about a real outward thing. 
I must either reduce it to the statement, ' certain sensations 
grouped under the designation ' gold ' have been constantly 
accompanied by that of yellowness,' or must take it to 
express the analysis of a 6 nominal essence, 5 — ' the concep- 
tion gold includes the attribute of yellowness.' 

37. If we could suppose Locke reading this chapter of 
Mill, he would say, If by a phenomenon you mean an idea 
(modification of consciousness) caused by an outward thing, 
or which gives any knowledge of an outward thing, I cannot 
admit that ' gold is yellow ' (which = c all gold is yellow ') 
represents a 'relation between phenomena,' or a fact 
relating to the outward thing, gold, and to the impression 
made by that outward thing upon human organs. To make 
it represent such a relation, you must reduce it to * this 
gold is now yellow ' (in the sense that something which now 
causes a sensation of hardness, and involuntarily recalls ideas 
of certain other sensations, also produces a sense of yellow- 
ness), or ' certain sensations of hardness have been constantly 
accompanied by that of yellowness.' And, further, if by 
c thing' you mean more than 'nominal essence,' more than a 
mere abstraction of the mind, you must admit that no pro- 
positions about things admit of the generality involved in the 
proposition ' gold is yellow.' 

Now there is no doubt that Mill is right in saying that 
such a proposition does express belief about ' things.' The 
question is whether his doctrine is compatible with that 
antithesis between things and thought which he retains, — 


with the doctrine that a thing, to be a real thing, must be 
something with the constitution of which thought has nothing 
to do. It is in order to carry out this doctrine that Locke 
reduces the real thing to a ' parcel of matter ' of which we 
only know, when we experience a sensation, that it must be 
then and there acting upon us in order to produce the sen- 
sation. (For the same notion of real thing in Mill see I. vi. 
§ 2.) In reducing ' real thing ' to such a ' parcel,' Locke 
had not in truth got rid, as he supposed, of ( creations of 
thought,' but he had made the real thing something of which 
general predication was impossible, for such predication does 
not express that present sensation which forms the only 
knowledge of the real thing we can have. He admits, it is 
true, a ' probability ' that a ' parcel of matter ' which has 
often caused a sensation of yellowness in immediate sequence 
upon one, e.g. of hardness, will continue to do so ; but in 
our absolute ignorance of the { parcel ' (an ignorance which 
we cannot suppose removed without infringing the antithesis 
between reality and the work of thought) this ' probability ' 
must reduce itself, as it did under Hume's treatment, to an 
involuntary expectation, a purely subjective tendency, the 
statement of which in a general proposition would not relate 
to any ' real outward thing.' 

38. Now Mill contents himself with saying that the 
judgment ' gold is yellow ' is not a mere analysis of nominal 
essence, but expresses belief in regard to an outward thing, 
without showing how, according to Locke's notion of reality, 
such belief can be justified. It may be said, indeed, that his 
language about ' things ' is an accommodation to popular 
usage, and that his doctrine about the proposition is more 
correctly expressed by the statement that it represents a 
relation between two phenomena. In fact, ' phenomenon ' 
in Mill's language always carries with it a reference to a 
6 thing ' which ' exhibits it ' ; but, waiving this, we are only 
driven by the alternative phrase upon a further difficulty. 
If ' gold is yellow ' expresses a uniform coexistence in nature 
between certain phenomena, how is knowledge of such 
uniformity possible ? The frequent concurrence, or close 
sequence upon each other, of certain appearances is one 
thing ; their coexistence in an order of nature is quite 
another. Strictly speaking, ' appearances ' cannot be said 
to coexist at all, but only qualities which appearances are 


taken to represent. ('Phenomena,' like Locke's ' ideas/ allows 
of convenient ambiguity between appearance and quality.) 
How, then, does frequent sequence of one appearance on 
another become either real coexistence of qualities in a 
thing or a real uniform sequence of phenomena in an order 
of nature ? Locke says straight away that it does not. A 
proposition which states a simple fact of sequence alone 
represents ' real essence ' : propositions which state coexist- 
ence of qualities or universal sequence can only relate to a 
nominal essence, to the content of a mere conception in my 
mind. In pronouncing otherwise, in saying that a certain 
experience justifies us in ascribing combustibility as a 
permanent attribute to the f thing ' called diamond (I. v. 
§ 2), Mill is ascribing reality to (or, better, constructing 
reality upon) a e conception of our mind,' the category or 
formal relation of thing and quality. He is doing what 
Locke a hundred and fifty years before had seen to be 
incompatible with the antithesis between reality and the 
work of thought, which yet Mill retains. He is recognising 
a real world, formed by the intellectual interpretation of 
feelings as representing a system of things. He is finding 
reality, not in the mere ' present sensation,' but in relations 
by which each sensation as it presents itself is determined, 
though they are themselves present only as thought about, 
and in the ideal thing, < gold,' ' diamond,' &c. which these 
relations combine to constitute. 

39. Our conclusion, then, is that Mill's account of the 
import of propositions is quite right so far as it means that 
the judgment ' gold is yellow ' is other than an analysis of 
nominal essence, other than a statement of a more simple 
idea 'yellow' contained in a more complex idea 'gold'; 
that it expresses a belief, and a true belief, about a real 
outward thing; but that he is wrong in not seeing that this 
thing is a thing which only exists for thought, and is ' out- 
ward ' only in the sense of being in space — itself a relation 
constituted by thought — not in the sense of being ' outside 
the mind,' which in fact is nonsense. That to which any- 
thing is outside must be in space — must itself be outside 
that which is outside it. We allow ourselves to talk of 
things as outside the mind, but we should think it queer to 
talk of mind as outside things. 


Thus the first ' category ' implied in Chapter V. is that of 
thing and quality. The same is implied in Chapters II. and 
III. The classification of names into (a) singular and general, 
(b) connotative and non-connotative, (c) concrete and abstract, 
clearly presupposes a proposition representing a conceived 
relation of thing and quality. 


(Mill, Book I. Chapter II.) 

40. The distinction between ' singular ' and ' general * 
names, as Mill gives it (I. ii. § 3), presupposes the proposi- 
tions (1) ' this is a man,' (2) ' this is John.' Except in 
relation to a proposition the distinction is unmeaning. 
' John ' by itself means nothing. ' Man ' by itself means 
something only because it is the symbol for a multitude of 
judgments in which qualities might be asserted of 'man' as 
subject. 'John 'is significant in such judgments as 'this 
is John ' or ' John has red hair.' The distinction of these 
judgments from such as 'man is a mammal' or 'this is 
man ' lies in the more complex determination by attributes 
of the object ' John ' than of the object ' man,' from which it 
results that only one individual object (in the fourth sense 
of ' individuality ' given above) l can be brought under the 
former conception of attributes, while many can be brought 
under the latter. Thus no account can be given of the 
distinction between ' singular ' and ' general ' names, which 
does not presuppose (a) propositions, (b) the relation of thing 
and quality as involved in the proposition. The true dis- 
tinction is that between singular and general propositions : — 

( , T , i -, i • x The common feature of all these is that 

m (h\ T? ■ t I L the subject of the proposition is an 

K) >< rf£ 18 - 1S * I individual, and an actual or possible 

(c) This is a stone. I ,. , r. . ', •■• * 

K ' ' object of intuition. 

, s -r ■, Here the subiect of the proposition is 

(2) ft SKV3ES mo tion. } -* - "^ « p-^ <*»»* °< 

(b) Heat is a mode of motion 


What is common to (1) and (2) is the thought of a thing 
(object) under qualities. ' John ' in (a) and (b) represents such 
an object as much as 'man.' The difference lies in the 
complexity of the determination by attributes. 

41. The objection to Mill's account of the distinction 

1 [Sec. 27.] 

NAMES. 203 

between ' general ' and ' singular ' names is (a) that it 
detaches names from propositions, and (b) does not look into 
the reason why the ' singular names ' of which he talks are 
applicable only to one individual, any more than he inquires 
what is meant by an individual. Thus c this stone,' ' the 
sun,' and i John ' are all alike singular names. But the 
individuality expressed by ' this stone ' (or • the present king 
of England ') = separateness in time. It does not express 
the individuality constituted by the peculiar attributes of 
the stone, of which, when I talk of * this stone,' I may be 
wholly ignorant. On the other hand, the individuality 
represented by * the sun ' and ' John ' (if this means a 
particular person) is individuality constituted by attributes. 
The object thus determined by attributes may be presented 
in intuition, but it is not such presentation that the singular 
name in this sense indicates. One sort of individuality, 
however, can as little as another be expressed without a 
proposition. Mill admits (I. ii. § 5) that a singular name 
which is not connotative is unmeaning. Still ' it shows what 
thing it is we are talking about,' though it does not ' tell us 
anything about it.' But if it has no meaning, how can it 
show what we are talking about ? If * John ' to the person 
I am talking with did not recall an object determined by 
certain qualities, there would be no good in talking to him 
about John. As used in every proposition in which I employ 
it, it is connotative ; the difference between it and ' gold ' is 
that by itself it does not, like ' gold,' involve a multitude of 
propositions. You cannot write it out into a multitude of 
propositions. But this means that by itself it is a mere 
sound — represents no mental act at all. 

42. Apart from propositions, then, the distinction be- 
tween 'general' and ' singular ' names is a distinction between 
names that have a meaning and those that have none. And 
the meaning of those that have meaning is always resoluble 
into propositions. Only in propositions has a singular name 
a significance, and these propositions already imply the con- 
ception (a) of an object under attributes, — though, if the 
subject or predicate is singular, the object is one of actual 
or possible intuition, — and (b) of a many in one. In the 
judgment ' this is John,' on occasion of a present sensation, 
there is a recognition of an object known under certain 
attributes, which is one in contrast with the many intuitions, 


this, that, and that, in which it is presented. There must 
have been a multitude of other intuitions in which the one 
John has been presented, as well as of intuitions in which he 
has not been presented, to give meaning to the this. Thus, 
though a singular proposition always implies the presentation 
to consciousness of a single object of intuition (and in this 
lies the distinction of singular from general propositions), yet 
this single object is. a conceived unity of an indefinite number 
of acts of feeling, and its name is predicable of an indefinite 
number of intuitions,— of presentations in space and time. 
My intuition (or perception) of John, like that of the sun 
or moon, is different every time I see him. But the single 
object I know as John is not. There is a liability to fallacy, 
then, in speaking of the single object as an object of intuition, 
unless we understand that something other than intuition is 
needed to constitute consciousness of the single object, viz. 
conception of identity. The object of intuition always = a 
conceived object presented under conditions of intuition. 
' John,' then, so far as it has meaning (so far as it indicates 
a single person conceived under attributes), does not differ 
from * the sun ' (which Mill distinguishes from it as con- 
notative from non-connotative). ' John ' is not predicable of 
more than one object, but no more is ' the sun.' If ' the 
sun ' is an object conceived under attributes, so is ' John,' so 
far as it has meaning. If there are many intuitions which 
are referred to — taken to represent — the one object 6 sun,' 
so there are many intuitions which are referred to the one 
object ' John.' l Just as we found that the object of intuition 
(consciousness of an individual object capable of distinct 
presentation in space or time) implied conception, so we find 
that the singular proposition implies conception, the thought 
of an object under relations which connect it with (are com- 
mon to it with) all other things, though they distinguish it 
from them. Mere intuition would not suffice to give a 
singular proposition. 

43. Thus, if every proposition, general in form, that is to 
be really true, should turn out to be a multitude of singular 
propositions written short (as Locke would have said, and 
as Mill says, II. iii. § 3, though he contradicts himself in 
III. ii. § 2), we still should not have got rid of conceptions 
at once really true and general. The question whether 
1 See General Introduction to Hume, § 80 ; vol. i. p. 64. 

NAMES. 205 

general propositions to be really true must be so reducible, 
corresponds to the question whether a conception to be 
really true must admit of being intuitionalised. In regard 
to conceptions, we found l the true account to be that, sup- 
posing them to purport to be of nature, they must relate to 
objects of intuition, — the relations determining the thought 
object must admit of being presented as relations between 
perceived or constructed objects. Such a perceived or con- 
structed object can be the subject of a singular proposition. 
So a general proposition about nature must be capable 
of being verified by an observation or experiment, of which 
the statement would be a singular proposition. But it 
does not follow because the proposition ' gold is soluble 
in aqua regia * must be capable of verification by an 
experiment in which ' this gold ' would be perceived to be so 
soluble, that the general proposition is merely short for a 
number of such perceptions. It represents a conceived 
relation between the properties of gold and those of aqua 
regia according to an order of nature, — a relation not con- 
tingent upon my perceiving it, and of which the existence 
is not any the more certain because I perceive it often. 

44. Mill's * connotative ' terms, like his e general,' pre- 
suppose the category of thing and quality and propositions 
resting on this formal conception, as he virtually admits. 
The difficulty is as to names which he says ' signify a subject 
only or an attribute only.' The former are e proper names,' 
and in the discussion on singular names have been shown 
either to mean nothing at all or to be connotative, — to imply 
the conception of a subject under attributes. To the person 
who uses them they are on every occasion on which (in every 
proposition in which) he uses them specially connotative, 
though taken by themselves they do not carry with them 
any indication of what this special connotation is. As for 
those which ' signify an attribute only,' they only seem to 
do so through detachment from a proposition. Such names 
arise from the act of the mind in regarding what it has first 
regarded as an attribute determining a subject, as in turn 
itself a subject determined by attributes. Thus a subject 
having been thought under an attribute in the judgment 
' this water is green,' the attribute under the designation 
* greenness ' becomes in turn a subject, — ' green is the com- 

1 [Above, sec. 28.] 


plementary colour to red.' If the relation of substance 
and attribute were something found ready-made in things, 
with the constitution of which thought had nothing to do, 
this substantiation of attributes would be unintelligible. In 
truth, not only is it true that apart from thought there are 
no things, but what we begin by reckoning individual sub- 
stances — separate things, yapKna, kclO* avrd, &c. — are for 
the most part temporary substantiations. The category of 
thing and quality has no final truth in regard to nature. 
Everything is a qualification of other things, a { retainer to 
other parts of nature.' In the organic body, it is true, the 
effects of ' other parts of nature ' combine to produce a 
result which cannot be resolved back into those effects and 
reacts upon external influences ; and in this sense such a 
body may be held to be, not merely for us but in itself, an 
individual substance. The so-called c thinking substance,' 
the human personality, is not properly a substance at all, 
being one with the eternal thought which is the source of all 

45. Mill's * abstract names ' = the latter sort of non- 
connotative names, those which signify ' an attribute only.' 

c Positive and negative names.' 'Negative names con- 
note the absence of an attribute,' — but there must be the 
conception of the attribute or there is no meaning in the 

* connotation of its absence.' Negation, except as supposing 
4 position,' is unmeaning, and every position implies nega- 
tion. The assertion that anything is white is a distinction 
of it from things that are not white, is a denial of whiteness 
in regard to them. Here again embarrassment arises 
through the detachment of names from propositions. 

* Positive and negative ' is properly a distinction of pro- 
positions. It is only because the name 'man' is sundry 
propositions Bvvdfisi, that it can be called ( positive,' or that 
any significant negation is formed by putting a t not ' to it. 
Instead of dividing names into ' positive and negative,' if we 
want, as a logician should, to reach the primary elements of 
thought, we should say that every judgment, and proposition 
as the expression of judgment, implies correlative position 
and negation, — the qualification of an object by distinction 
from other objects, which amounts to the denial in regard 
to it of what belongs to them, and in regard to them of 
what belongs to it. c Omnis determinatio est negatio ; ' 

* omnis negatio est determinatio.' 



(Mill, Book I. Chapter III.) 

46. Thus from Mill's account of names we elicit the 
formal conceptions of one in many and many in one, of 
thing and quality, of qualification as a position which implies 
negation. These, so far, will be our ' categories,' — relations 
of which all knowledge is the further articulation, — the 
import of all possible propositions. Actual propositions 
mean much more than this, but they mean at least this, and 
only through conveying this can they convey more. 

Mill himself regards ' categories ' in the scholastic way, 
as ' summa genera ' of things. Eoughly one may say that 
there is a wrong view of the categories and a right one. 
The right one regards them as the relations or formal con- 
ceptions (which comes to the same, since conception consti- 
tutes relation), without which there would be no knowledge 
and no objective world to be known. They are not the end 
but the beginning of knowledge, not ultimate truths, but 
truths which we already know in knowing anything, though 
the correct disentanglement of them is in one sense the 
great problem of philosophy, — in the sense that on the 
correct solution of it the correctness of metaphysical philo- 
sophy in general depends. The wrong view goes along with 
the false notion that the essential of thought is abstraction, 
and that thus the end of thinking is to reach certain yevrj 
twv 6vt(0v, having no common elements with' each other. 
The knowledge of these, since more thinking means more 
knowing, must be the highest knowledge ; they are yvcopi- 
ficoTara aifKws. According to one view they are things ; 
according to the other relations. According to one they are 
the end of knowledge ; according to the other, the beginning. 
According to one they are really apart from the objects of 
ordinary knowledge and experience, and are known by ab- 
straction from these ; according to the other, all objects of 
ordinary knowledge and experience are determinations of 


them, so that we know them in knowing the former, though 
we do not know that we know them. In Aristotle both views 
appear. On the one hand, they are a^fjuara ttjs Karrjyoplas. 
The classification of them is the classification of the ' import 
of propositions.' They represent different meanings of \<m 
in the judgment, different modes of the relation which every 
proposition represents. The actual list of them confirms this 
interpretation. Whatever fault may be found with it, it is 
not a classification of things, but of relations between things. 
This antithesis is not to be understood as meaning that 
there really are things apart from relations. What is 
meant is, that admitting the vulgar view of things, as some- 
thing in themselves apart from relation, Aristotle's classifica- 
tion is not to be taken as a classification of such things. On 
the other hand, he speaks of them as ysvy twv ovtoov, of 
which ' first philosophy ' is the knowledge. 

47. Mill professes to give a classification of existences. 
Such a classification cannot serve the purpose which a 
theory of the categories should serve, — that of disentangling 
the presuppositions of knowledge and experience, — for it 
necessarily involves them. It is in fact a theory of the 
universe, and, whether a right or wrong theory, does not 
supply but presupposes the answer to the question, how 
such a theory is possible. Thought brings with it, to the 
classification, itself and the relations which it constitutes in 
the act of knowing. Thus, in order that I may know ' states 
of consciousness ' as c existences,' I must already have con- 
verted * feeling ' into ' felt things ' ; i.e. by an act which 
constitutes the category or relation of identity (one in many), 
I must have converted the feeling into an object which 
remains after the moment of feeling is over, and is one and 
the same in the manifold recurring instances of its being 

48. Further, according to Mill (I. iii. § 3), a ' state of 
consciousness ' has to be distinguished, if a thought, from 
the object thought about, if a sensation, alike from the 
c object which causes it ' and from the ' attribute which we 
ascribe to the object in consequence of its exciting the 
sensation.' Such a distinction implies a further exercise of 
the same intellectual act (the same ' function of judgment,' 
in Kant's language) by which a feeling became a knowable 
object. It implies not merely that a feeling is objectified, 


but that the objectified feeling is distinguished from a thing 
which causes it and which it represents, — a thing determined 
by properties which are its content, as distinct from the mode 
or form of its presentation to consciousness. Thus ' the sun 
is not my idea of the sun,' and any sensation I receive from 
it is other than the property of exciting that sensation as it 
exists in the sun. In order that such distinction may be 
possible, formative conceptions not only of thing and quality 
but of cause and effect must be presupposed, — conceptions 
which cannot properly be regarded as the result of expe- 
rience, because without their operation the succession of 
feelings could not become that experience of things, of a 
connected objective world, from which alone — as opposed to 
the mere succession of feelings — the doctrine which ascribes 
them to experience seeks to derive them. 

49. Having thus already implicitly assumed the cate- 
gories of cause and substance in his account of the first sort 
of nameable things, viz. states of consciousness, Mill pro- 
ceeds explicitly to introduce them as two sepai^ate sorts of 
'nameable things,' under the designation of (a) the 'hidden 
external cause to which we refer our sensations ' (I. iii. § 8), 
and (b) ' something distinct from sensations, thoughts, 
&c. . . . the being that has the thoughts' (ib.). Not 
having got hold of the true view of categories as conceived 
relations, — relations constituted in and by the act of concep 
tion or knowing, — he does not see (a) that the f. unknown 
external cause ' and the ' unknown inner substance ' are 
each merely one member of the relation which thought 
constitutes in the very act of knowing a ' state of conscious- 
ness,' and (b) one and the same member. We have shown 
how the relation of one in many and many in one is 
involved in this act, and how, so soon as the known state of 
consciousness is known as representing a thing in which 
qualities corresponding to or causing the state of conscious- 
ness exist, the relations of substance and attribute, cause and 
effect, are involved too. What Mill does is to detach one 
member of each of these kindred and mutually connected 
relations, — the one from the many, the substance from the 
attributes, cause from the effects, — and treat it as a separate 
thing, which is really to make it nothing. Thus he gets a 
'thing in itself as the unknown single cause of manifold 
effects in the way of sensations, existing in abstraction from 
vol. n. p 


these effects, and an unknown self or mind to which these 
same sensations (and thoughts, which according to him are 
their products) belong, as existing in abstraction from them. 
He does not see that either in such abstraction mind and 
thing are nothing, or that rightly understood the two ( nou- 
mena' are one and the same member of the formal rela- 
tions mentioned. Mill dexterously avoids using the same 
language in regard to the ' thinking thing ' and c body,' but, 
when we look closely, we find that in effect what is said of 
one is said also of the other. ' Body ' is spoken of as cause, 
' mind ' rather as substance. Body causes sensations, ideas 
belong to the mind. But then we find that, though sensa- 
tions are not, as sensations, attributes of body, yet that 
body, if it is to be anything at all (what Mill calls an 
' objective fact ' : see end of Chapter III.), must be regarded 
as having attributes which cause the sensations, and thus as 
a substance. Again, the moment we regard mind as a sub- 
stance related to ideas as attributes, — the moment we get 
over the supposition that as a substance it has any reality 
apart from its attributes, — we regard it as manifesting itself 
in these ideas, i.e. as a cause. ' But at any rate (it may be 
said) 'body' is external, the * thinking thing' is not.' 
We cannot, however, make body external to mind, without 
making mind external to body. 

In short, ' body in itself ' and ' thinking thing in itself ' 
are alike unmeaning abstractions of one factor in a logical 
category from the other factor. As soon as we try to give 
meaning to either by restoring the other factor in the 
category, we find them equivalent. Each becomes the cate- 
gory of one and many, substance and attribute, cause and 
effect, outward and inward. 1 

50. Undoubtedly there is a difference between sensation 
and quality of body causing sensation, between the sensation 
of light and the undulations of aether, affecting certain 
nerves in a certain way, which cause it. This, however, is 
not a difference between the operations of an unknown body 
on one side and the attributes of an unknown mind on the 
other, but a difference between, and at the same time a 
correlation of, a known sensitive organism on the one side 
and known non-sensitive bodies on the other. This is an 
ascertained physical distinction which presupposes logical 

1 [' Outward and inward ' is queried in the MS.] 


categories to make it possible, but is quite other than they. 
With this we confuse the unreal opposition between body in 
itself and mind in itself, of which in truth each is the same 
with the other, each being one member of the logical 
categories above explained in imaginary abstraction from its 

Thus, when Mill says (I. iii. § 8), ( But of the nature of 
body or mind, further than the feelings which the former 
excites and which the latter experiences, we do not, accord- 
ing to the best existing doctrine, know anything,' if ' body ' 
and * mind ' mean body as the cause of sensations and mind 
as the sensitive organism, we do know a great deal about 
them. It has been ascertained with great clearness what 
specific sensations are related to what properties of bodies. 
We know the exact sort of vibration which excites each 
minutely different sense of sound. If by ' body * and ' mind ' 
we mean two ultimate causes or substances in supposed 
abstraction from all effects and attributes, one outward the 
other inward, we know all about them too, for we know 
that they are merely different expressions for one and the 
same logical category, misunderstood. The only ' thing in 
itself ' is the thinking subject, which is not cause or 
substance, but the source of the categories of cause and 

51. We find, then, that t states of consciousness ' are 
wrongly put as a class of existences alongside of substances, 
whether ' bodies ■ or ' minds,' as another sort of existence, 
and ' attributes ' as a third, since * substance ' is but one 
member of a relation (of which ' attribute ' is the comple- 
ment) involved in the act by which a state of consciousness 
becomes an object of knowledge. Not only so : since the 
recognition of objects of knowledge as in any sense ' founded 
on states of consciousness' (that they = states of conscious- 
ness is untrue, as will appear) at any rate comes quite late 
in the process of knowledge, it is a mistake to include such 
states among the categories which, according to the true 
view, purport to be the conceived relations involved in all 

52. If it is asked, Why assume that this is the true view 
of categories 9 Why should not a ' classification of name- 
able things ' after MilPs manner take its place ? the answer 
is, (a) you cannot make such a classification without the 

p 2 


presupposition, whether acknowledged or not, of these 
conceived relations, — as we have found to be the case with 
Mill ; and (b) any classification, save of formal conceptions, 
cannot be exhaustive, e.g. such a one as that into which 
* states of consciousness ' could legitimately enter. We can 
make it exhaustive, indeed, like any classification, by 
dichotomy, but this is unmeaning. * States of consciousness ' 
only properly enter into a classification which would also 
include (a) manifold forms of unconscious existence, (b) 
manifold modes of consciousness, and such a classification 
could not be exhaustive. Mill sometimes writes as if every- 
thing could be reduced to a state of consciousness. This is 
his form of Berkeleian and Humian idealism. He admits the 
contrary, however, virtually in I. iii. § 4. That everything 
is a state of consciousness is the false idealism ; that apart 
from, or except in relation to, a thinking consciousness 
there is no reality or existence, is the true, which is quite 
compatible with the admission of any amount of forms of 
unconscious existence. Mill, so far as he can, maintains the 
former false idealism, but is constantly obliged to drop it. 
He never gets hold of the latter. 

53. As if he had not already dealt with ' attributes ' in 
dealing with ( substances ' as causes or subjects of sensations, 
Mill (I. iii. § 9) goes on to a separate inquiry about them. 
' The distinction which we make between the properties of 
things and the sensations we receive from them ' he calls 
merely ' verbal.' Doubtless, if ' sensation ' means the recog- 
nition of a feeling as received from a thing, it only differs 
' verbally ' from the recognition of a property of a thing, in 
the sense that in such recognition of a feeling I think of a 
relation between a conscious subject and a thing which 
amounts to a determination or attribute of the thing. The 
judgment (a) ' the thing is white,' and the judgment (b) ' I 
receive a sensation of whiteness from the thing ' are equiva- 
lent in virtue of the common category of thing and quality 
involved in each. The mere sensation is one judgment as 
little as the other, and we have no reason to suppose that 
the animals, though undoubtedly they have the sensation, 
ever form either of these judgments. From the apparent 
absence of language among them we infer that they have no 
need of it because they do not convert mere feeling into a 
felt thing. But, though the two judgments involve the same 


category, it is a mistake to treat them as equivalent ; still 
more to try to reduce (a) to (b). (b) really means more than 
(a), and is psychologically later than it. It is in virtue 
indeed of an intellectual interpretation of feeling that we 
ascribe qualities to things ; but this ascription takes place 
before we reach that clear conception of the distinction 
between the thing affecting and the subject affected which 
is implied in (&). And when this is reached, and though we 
may be quite aware that whiteness is no c entity per se called 
a quality' (I. iii. § 9), but implies relation to a particular 
sensitive organism, we mean different things by (a) and (6). 
(a) implies belief as to a nature, a permanent order of things, 
— a belief, doubtless, accompanied in the instructed person 
by knowledge that his sensitive organism is a factor in this 
order ; (6), by laying stress on the sensitive organism, 
naturally conveys the belief that the present experience is 
due to some peculiar condition of this. Mill, ignoring the 
difference between feeling proper and the intellectual act 
through which feeling becomes a known fact, — a sensation 
which I regard myself as receiving from a thing, and which 
thus becomes a determination alike of me and of the thing, 
— looks on all attributes in the way of quality simply as 
feelings. He treats attributes in the way of quantity in the 
same way. ( Nobody, I presume, will say that to see, or to 
lift, or to drink ten gallons of water does not include in 
itself a different set of sensations from those of seeing, lifting, 
or drinking one gallon ; or that to see or handle a foot-rule, 
and to see or handle a yard measure made exactly like it, 
are the same sensations ' (I. iii. § 12). Therefore quantities 
as well as qualities are included under the category ' states 
of consciousness.' 

54. Attributes in the way of relation, however, require a 
class to themselves. Ordinary attributes, such as are pre- 
dicated in the judgment, e this snow is white,' he does not 
treat as resting on a relation at all, though according to his 
account of it it expresses consciousness of a relation between 
the sensitive organism and the exciting object, and according 
to any other possible account of it it implies a relation 
in the way of distinctness between one colour and others. 
He treats it, further, as if it represented a single sensation, 
not seeing that the relation of identity between a present 
and numberless past experiences is implied in the recognition 

214 LOGIC : MILL. 

of ' this ' as snow. Nor does the attribute of quantity strike 
him any more as implying relation. He seems indeed to 
allow that there must be a set of sensations to constitute a 
quantity, but he takes no notice of any relation between 
them. He does not consider, e.g. whether a coloured surface 
would form or be a magnitude but for the relation of co- 
existence between the several pictures which its parts present 
to me : whether ' ten gallons ' has any meaning except as 
determined by relation to nine gallons, eight gallons, &c. 
The only attributes which he regards as attributes in the way of 
relation are those indicated by names which imply a correlate. 

Take the judgments, (1) ' this weighs ten pounds,' (2) 
' snow is white,' (3) ■ snow falls after a frost, (4) ' snow is 
like wool.' Mill would say that whereas only one pheno- 
menon is represented by (1) and (2), (or, if more than one 
by (1), yet the several phenomena are somehow formed into 
one * set '), (3) and (4) express two phenomena connected in 
a particular way. In fact, however, a relation between a multi- 
plicity of objects is equally implied in (1) and (2) ; only in 
them the other objects by relation to which the given object 
is determined are not specified. In other words, the relations 
expressed in (3) and (4) are more determinate and specific 
than those expressed by (1) and (2). 

55. Mill distinguishes two kinds of relations, (a) those 
' grounded on a series of states of consciousness,' in the 
production of which the related objects jointly take part, (b) 
those which are not grounded on states of consciousness, but 
are themselves states of consciousness, such as order in place, 
order in time, and resemblance. His account of (a) only 
differs from his account of attributes in the way of quality 
in this, that when a state of consciousness or series of such is 
produced by one object, an attribute grounded on it is a 
quality, when by more than one, a relation. This conversion 
of sensation into an attribute ' grounded on it ' presupposes, 
as we have seen, conceived relations. The true account of 
the difference between the two cases is that in one the pre- 
supposed relation or conception is that of thing and quality 
simply, in the other the quality is unfolded into a relation 
between the subject-thing and other things, between the 
things which, as Mill puts it, jointly produce the state of 
consciousness. In no case can the object to which an 
appearance is referred as a qualification of it or as an instance 


of it, be ultimately thought of save as conditioned by other 
objects, as deriving its essence from relation to them. But 
in some cases not merely the essence of an object, but its 
existence, implies relation to another; and in such cases 
(perhaps in others) the correlation of the objects has been 
marked in language, ' father and son,' &c. Such conditions, 
implicitly conveyed in such a predicate as e white,' are explicit 
in a predicate which contains a preposition, e.g. ' this is a 
man on horseback,' ' snow falls after frost.' 

56. (&) Relations of resemblance and succession, according 
to Mill, are states of consciousness made up of other states, 
certain states of consciousness put together. The effort of 
all idealism of that sort which takes idea as = feeling must 
be to reduce ' objects ' to feelings and compounds of feeling. 
Once admit relations other than such compounds, and 
you have either to recognise an order presented to us from 
without other than that constituted by feelings or ideas, or 
to admit that something results from the action of thought 
on feelings other than the feelings, which means that thought 
originates (gets something from feelings not given in them). 
Hence Mill instinctively tries to reduce ' objects and attri- 
butes ' to states of consciousness, and ' relations ' to com- 
pounds of such. Resemblance according to him = feeling of 
resemblance, and that not a third feeling over and above 
the two resembling feelings, but these two together. Three 
objections may be made to this; (1) How do successive 
feelings put themselves together ? (2) The relation of resem- 
blance between two feelings of colour must on this view itself 
be a colour. (3) As the feeling of colour which I had 
yesterday is past, never to return, and likewise the feeling of 
today, how is it that the likeness between them, which is 
merely the two together, remains to be talked and thought 
of ? The same holds, mutatis mutandis, of time. In regard 
to all relations, then, we must hold that they are not states 
ofconsciousness. That they are grounded on such we may 
admit, in the sense that if we did not feel there would for us 
be no world of related objects, though we might feel without 
there being for us any such world. 

57. It is from treating names apart from propositions 
Jiat Mill fails to see that all names, as significant, or as 
used in predication, are of relations. When he comes to 
treat of propositions he virtually admits this. In I. v. § 6 


e.g. lie says that the proposition, ' the colour I saw yesterday- 
was a white colour,' involves an assertion of resemblance. 
Clearly, on the same principle, ' this is white,' ' snow is 
white,' involve the assertion of resemblance. The relation of 
identity involved in all judgments Mill ignores. ' This is the 
same colour I saw yesterday ' means, according to him, ' very 
like it,' for ' the feeling which I had yesterday is gone never 
to return ' (I. ii. § 11). It is quite a different sense, according 
to him, in which we are said to be ' sitting at the same table.' 
But if you reduce the table to ' states of consciousness ' in the 
sense of feelings, you have a like difficulty about the identity 
of the table. The table cannot be reduced to a single state 
of consciousness, but only to several, each of which is gone 
f never to return ' before another begins. In truth, the 
table is not any series of states of consciousness, but an 
object constituted by an intellectual synthesis of these. So 
y colour' in the above judgment is an object constituted by 
an intellectual synthesis of feelings, to which in the above 
judgment I refer a present experience as an instance of it. 
The distinctness in time between the present and previous 
feelings no more interferes with their understood identity 
than does that between the states of consciousness in which 
the table is presented with its identity. The apparent 
difference arises from the fact that different parts of space 
coexist, while different times do not. Thus it seems that, 
while the feeling of today cannot be the same with that of 
yesterday, the part of the table at which A sits can form one 
whole with that at which B sits. What is overlooked is that 
it is only because the part of the table where A sits and that 
where B sits are each other than my state of consciousness 
that they can coexist. Certain states of consciousness must, 
in virtue of the equal presence of the ego to them, have been 
converted into coexisting parts of a whole, before I can con- 
ceive of one table at several parts of which A and B sit. 
And a like operation of thought on successive feelings of 
colour converts them into an identical object. 

58. To sum up : the best view of that which the propo- 
sition expresses is that it is \ the thought of an object under 
relations.' This is equally true whether the object is ' given 
in intuition,' and thus the judgment is singular, or no ; 
whether it is merely conceived or also perceived or imagined. 
The most primary of these relations (those involved in all 


judgments) are (a) identity, which implies difference ; ' this 
is a fire ' ; (b) thing and its action (cause and effect) ; ' this 
warms me ' ; (c) thing and quality ; ' this is hot ' ; (d) de- 
termination of one object by another, which is also their 
reciprocal action ; ' this which is warm is other than that 
which is not.' 

59. Without a feeling consciousness of likeness and un- 
likeness, it is true, none of these judgments would be 
possible ; but I should not introduce consciousness of like- 
ness and unlikeness among the categories or relations 
involved in all judgment, because consciousness of likeness 
and unlikeness is possible without any judgment, possible 
in a merely feeling subject, in the sense that an animal could 
not e.g. be taken out of hot water and put into cold without 
being conscious that the latter feeling was unlike the former. 
This, however, would not amount to consciousness of a 
relation of likeness and unlikeness. 1 In order to this there 
must previously be a conception of objects, permanent in 
distinction from transitory feelings, between which such a 
relation may obtain. In such a conception of objects the 
above categories are involved. There can be no conception 
or judgment (which is but the evolution of conception) at all 
without them, 2 but there can be without a conception of 
relation in the way of likeness or unlikeness (which pre- 
supposes that of identity and difference ; ' these colours are 
not the same, but like ' ; ' these colours are not merely differ- 
ent, but unlike'). 

Thus there may be consciousness of likeness and unlike- 
ness in one sense without judgment at all. In the other 
sense of such consciousness there may be judgment without 

1 In the case of a cat thrown out of objects, — does not retain its feelings as 

warm water into cold, the latter feeling, objects still there for thought when 

as felt by the cat (whom we suppose they have ceased to be felt, and for the 

not to be a thinking subject), is deter- same reason is not conscious of a rela- 

mined by contrast to the former. Its tion of unlikeness as a relation ; which 

painfulness (the shock caused) just implies coexistence of the related ob- 

consists in this contrast. There must jects. We cannot speak of the animal's 

then be continuity of feeling conscious- consciousness except in terms of our 

ness from one experience into the other; own, and hence call its feeling of cold, 

one feeling soul equally present to both as determined by the previous feeling 

events. But a feeling soul is not of warmth, a consciousness of relation, 

therefore a thinking (self-conscious) which it is not. 

soul. It does not separate itself from 2 Can the above categories be made 

itself and present itself to itself as an out to be involved in such a judgment 

object. Hence it does not present its as 'something is here'? 
feelings to itself as permanent felt 


it. Nor is the statement of resemblance ever the whole 
import of a proposition. A proposition which asserts re- 
semblance between phenomena always asserts it as a quality 
of a thing. c The sun is like a ball of fire ' ; what is here 
asserted is a likeness between certain appearances as a quality 
of the sun. So with such relations as order in place and 
time ; they stand on a different footing from the above. 
They are not involved in all judgment, and in every judg- 
ment which states them they qualify a subject and thus pre- 
suppose the other relations. 

60. Their predominance in Mill's account arises from 
his effort to make all judgments relate to states of conscious- 
ness. 1 Two really inconsistent views of the proposition run 
through Mill, one according to which it states a relation of 
attributes, the other according to which it states a relation 
between phenomena, which he interprets as = states of con- 
sciousness, = feelings. The truth of the matter is not either 
that some propositions express a relation between attributes, 
others sequence or coexistence between phenomena, or 
that the relation between attributes is reducible to sequence 
and coexistence between phenomena, but that every propo- 
sition expresses a relation between attributes or a deter- 
mination of an object by attributes, and that the attribute 
by which the object is determined, or which is asserted to 
coexist with other attributes, sometimes consists in the 
sequence or coexistence of phenomena. 

61. We must observe in the first place that if phenomena 
= mere feelings, as they are felt, not as they are thought, 
no relation between them is possible. The members of a 
relation must exist together. But of the feelings between 
which we assert a relation one is past or passing before the 
other begins, and this other has no sooner begun than it is 
over. I strike one note of music and then another and 
assert a relation of difference between them, but only because 
for the comparing subject they are present together, — only 
because for it they are not what as feelings in time they 
are, viz. successive. It may be said, ' Are not the feelings 
of animals related ? ' We answer, really they are, but not for 
the consciousness of the animals. Phenomena, then, in 
order that relations of sequence and simultaneity may obtain 
between them, must be, not feelings as merely felt, but felt 

1 See passages in I. t. §§ 4 and 6. 


objects, feelings as thought of. Secondly, when such sequence 
and simultaneity of phenomena form part of a judgment, 
by a mental act, the same in principle with that by which 
feelings become felt things, the relation in question becomes 
an attribute or determination of a thing. A judgment, if it 
is to state a relation between attributes, cannot state merely 
a sequence or simultaneity of feelings ; it must state the 
sequence of feelings as the attribute of something. ' Rain 
falls after a lessening of atmospherical pressure,' 6 rain falls 
when heavy clouds are overhead.' Here no doubt in the 
several cases a sequence and simultaneity of phenomena are 
stated, and also a relation of attributes is stated, but the 
phenomena between which sequence or simultaneity is stated 
are not the attributes between which a relation is stated. 
The phenomena are the fall of rain and the lightening of 
atmospheric pressure or the appearance of clouds. The 
attributes between which the relation is stated are on the 
one hand general attributes of rain, on the other the 
attribute consisting in the sequence of its fall upon light- 
ening of pressure or its simultaneity with the appearance of 

62. We must not, then, use relation of attributes and 
relation of phenomena as equivalent expressions. If we take 
the former as expressing the import of propositions, we can 
only say that in some cases the attribute predicated consists in 
the simultaneity or sequence of phenomena, but the relation 
between attributes is never one (a) of sequence, or (b) of 
simultaneity, or (c) of mutual limitation in space, (a) and (6) 
are relations of events, and attributes are not events, though 
an attribute, in Mill's language, may be ' grounded on ' an 
event. ' John's hair turned grey before he cut his wisdom 
teeth ' ; here are two sequent events stated, but the two 
sequent events are not two sequent attributes. The attribute 
consists in the relation between the events and is an attribute 
of John, with whose other attributes, if you like, it co- 
exists, but not in the way of cotemporaneity, nor in the way 
of mutual externality. Mutual externality of parts or ' order 
in place ' may be asserted as an attribute (' bodies are ex- 
tended '), but this attribute does not therefore coexist in the 
way of such externality with that of which it is the attribute. 
Attributes can no more be external to attributes, or to that 
of which they are attributes, than thoughts to thoughts, or 

220 LOGIC : MILL. 

than the conception of space can itself be a space. Take 
the propositions about triangles in Euclid. They assert 
attributes of triangles. These all represent relations of 
spaces each outside the other, but the several attributes are 
not therefore outside each other nor outside the triangle. 
The triangle, as intuited, is so much space, but not as con- 
ceived, and it is as conceived that it is a subject of attributes. 
As conceived, the relation of externality has no application 
to it ; its attributes are not external to it or each other ; 
there is no relation in the way of order of place between 
them and it or each other. If such a relation cannot obtain 
even between the attributes of a triangle, there is even less 
excuse for supposing it to obtain between the attributes of 
anything else. 



(Mill, Book I. Chapter VI.) 

63. Mill's distinction of real and verbal propositions 
corresponds to Locke's between 'instructive' and 'trifling.' 
The latter are those in which the predicate merely represents 
an idea already included in the complex idea which the sub- 
ject-name stands for (cf. Mill, I. vi. § 4). It is difficult, 
however, to ascertain what propositions are to be so regarded. 
From § 5 it would seem that ' all men are mortal ' is not 
verbal. ' All men are animals ' presumably would be. Yet 
it is difficult to allege that ' mortality ' is not just as much 
asserted of a subject in calling it ' man ' as is ' animality.' 
All propositions, according to Mill, are merely verbal 
which are ' essential ' in the Scholastic sense, since ' essence ' 
= the meaning of a name. There is no difficulty in deciding 
what attributes are of the essence, so long as you regard 
essence in the Scholastic way as fixed by a definition ; but 
once admit that the meanings of names fluctuate, that the 
same name means different things to different persons, and 
it becomes impossible to say, in regard to any general pro- 
position, unless we maintain (which Mill does not) that the 
mere fact of its generality proves it verbal, whether in the 
predicate you assert more than you imply of everything to 
which you apply the subject-name. The formal logician (in 
the special sense), it is clear, can have nothing to say to it un- 
less it is merely verbal. Unless the complex idea expressed by 
the name 'man' includes the idea of mortality, you must ' go 
beyond the given conception,' — you must refer to ' matter' not 
included in it, so that your thinking will be other than for- 
mal, — in the syllogism which proves mortality of 'this man.' 

64. According to Locke's own doctrine, again, there is 
no doubt what general propositions are 'trifling,' what 
' instructive.' All general propositions are ' trifling,' unless 
they relate to morals or mathematics. According to the 

222 LOGIC : MILL. 

predominant notion of the ' real ' in Locke, as ' that which 
happens to me whether I will or no ' in opposition to all work 
of thought, a proposition which is to state real truth should 
state a single event in the way of feeling. It has been 
sufficiently shown that a proposition stating merely a present 
feeling apart from modification by what is not present feeling 
is in fact impossible. Locke, however, so far tries to be con- 
sistent that (subject to exceptions in favour of mathematics 
and morals) he restricts real propositions (statements of real 
truth) to the singular form. According to him they must 
state an event now happening or that did once happen 
(though he cannot in effect keep such events clear of modifi- 
cation by understood relations) . A ' real coexistence of 
ideas ' (which is about equivalent to a ' simultaneity of phe- 
nomena ' in later language) can be asserted in present or past 
single instances, but not generally. When it seems to be 
asserted generally ('all gold is soluble in aqua regia '), what 
is in fact asserted is either a ' fact in my mental history ' (to 
use Mill's phrase), and that a single fact, or that a certain 
idea, consisting in a strong expectation, is included in the 
6 nominal essence ' of the subject. Of ' mixed modes,' how- 
ever, with Locke the ' nominal essence ' is also ' real.' The 
ideas of them are our own making, not made for us by some- 
thing without us, and do not, like those of substances, relate 
to an archetype other than themselves. The certainty of 
propositions about them, accordingly, may be both real and 
general. According to the original account of modes in 
chap, xii., &c. of Book II this could only be consistently 
admitted of 'mixed,' not of 'simple,' modes, for ideas of 
simple modes of space and duration, according to this 
account, are just as much made for us as simple ideas, and 
have just as much reference to outward things as have ideas 
of substances. However, in the third and fourth books he 
extends to mathematical ideas (which are in fact ideas of 
simple modes) as well as to moral ones the privilege of having 
no relation to reality other than themselves. In regard to 
mathematics and morals, then, we may extend our knowledge 
without experiment and observation, — without new events 
in the way of sensation happening to us, — and can arrive at 
propositions new and ' instructive ' (in the sense that what 
they state in the predicate is not already implied in our 
having the idea which forms the subject), and universally true 


(because they state relations of ideas which belong merely to 
our minds, and which, therefore, must be always what they 
are once), by some process of * laying in order our ideas.' 

65. Locke leaves us quite in the dark as to the nature of 
this process, or how it is possible — since it implies some 
construction of new ideas — for a mind which can only com- 
bine and abstract given ideas. What forced such admissions 
upon him was, no doubt, the existence of the sciences of 
jurisprudence and geometry, both of which seem to arrive at 
new truths without experiment or observation, and to argue 
deductively from these as universally true, without liability 
to exception or modification. In the fifth proposition of 
Euclid, having got a certain construction, we are able by a 
series of direct comparisons of magnitude to compare the 
angles of the isosceles as remainders of equal angles and 
therefore themselves equal. As Locke would say, we arrive 
at the conclusion by ' laying in order intermediate ideas/ 
But having arrived at it by means of a particular construc- 
tion in that ' bare instance,' we take it as universally true. 
Afterwards we try to exhibit angles as angles at the base of 
an isosceles, in order at once to be sure of their equality. 
Locke explains this by saying in effect that the ideas 
between which equality is demonstrated in the single case, 
being merely our own making, do not depend on anything 
else which can modify them. So in jurisprudence new 
general propositions are made : a general definition of a 
crime is constructed (representing no doubt a prior custom, 
but then this custom is man's making), and then the 
business of the jurist is to exhibit particular cases in such a 
way as that they may be covered by the definition, just as 
the geometer tries to find a construction by which (say) he 
may exhibit a couple of angles as angles at the base of an 
isosceles. In regard to jurisprudence it is clear that the 
difficulty of admitting such certainty, at once real and 
general, is not the same as in regard to mathematics. Its 
general propositions, the f empiric ' will say, represent 
human convention or enactment. Rights having been 
admitted and acts made criminal, as occasion required, 
under direction of a common understanding, but without 
explicit recognition of a general principle, the jurist reduces 
the rights or crimes in question to such a principle, and 
then brings fresh cases as they occur under such principle. 

224 LOGIC : MILL. 

In all this he is merely summing up and finding general 
expression for what does not profess to be other than the 
work of man. (Convention and enactment, however, imply 
that man is able to make new conditions for himself in a 
way of which animals are incapable, and which would be 
unaccountable if human life were simply the result of 
natural influences : they imply that man presents himself to 
himself as the possible subject of a condition other than 
his present one.) Mathematical propositions, on the other 
hand, relate to what is just as much natural as anything 
can be, nay, to what is natural /car i^o^ijv, to elementary 
properties of matter. In Locke himself we find that ideas 
which in Book IV are treated as merely of the mind, have 
in Book II been treated as ' ideas of primary qualities of 
matter,' of qualities which * are in things whether we per- 
ceive them or no.' l 

66. Kant retains from the school of Locke the doctrine 
that judgments of ' empirical origin,' if more than analyses 
of a complex idea, cannot be universal and necessary. This 
is indisputable, if by such judgments is meant a statement 
of an event in the way of feeling, or a summary of such 
events, and this is what it did mean to Kant's predecessors, 
and what alone it can properly mean unless experience is to 
be recognised as in its very origin determined by thought. 
If a judgment, general in form, is a summary of past events 
which have hitherto happened without exception in a 
particular way, it does not purport to be universal and 
necessary. If it is taken to represent the strongest possible 
subjective expectation, this is relative to the individual's 
experience so far as it has gone. People who think that 
the development of habits through hereditary transmission 
will account for the necessity of necessary truths, show 
that they do not know what is meant by such 'necessity.' 
It does not mean strength of subjective expectation, which 
is what alone could be thus accounted for, but necessity as 
a condition of there being a nature. 

67. The propositions which Locke spoke of as explana- 
tions of a ' nominal essence,' Kant called f analytical 
judgments.' All universal judgments, then, if of empirical 
origin, must be analytical. If there are judgments which 

1 See General Introduction to Hume, §§ 115 if. ; vol. i. p. 94 ff. 


are universal and necessary, and at the same time synthetical 
(' adding to our conception of the subject a predicate not 
contained in it'), they cannot be of ' empirical origin,' they 
must be ' a priori. 9 Kant finds such judgments in mathe- 
matics and natural philosophy. His instances are, i 7 + 5 
= 12 ; ' c a straight line between two points is the shortest ; ' 
c in all changes of the material world the quantity of matter 
remains unchanged ; ' 'in all communication of motion 
action and reaction must always be equal.' These are not 
summaries of events in the way of feeling — not in that sense 
results of experience, though without experience we could not 
have them. They arise out of the presuppositions of experi- 
ence, out of conditions under which alone experience is 
possible; the two former out of conditions under which 
alone intuition of objects is possible (presentation of them as 
individual in space and time) ; the two latter out of those 
under which alone experience as of nature, of an objective 
uniform system, is possible. 

68. Kant's distinction, however, of ' synthetical ' and 
' analytical ' propositions, as he understood it, does not 
exactly square with Locke's of ' instructive ' and ' trifling,' 
as this was understood by Locke. Kant would not have 
admitted that his analytical propositions were merely verbal 
or trifling, or that the conception analysed in such a propo- 
sition could be no more than a combination of ' ideas ' 
derived from previous experience and fixed by a name. Nor 
by 'judgments of experience which, as such, are always 
synthetical,' did Kant understand merely summaries of 
events in the way of feeling, undetermined by conceptions 
not derived from such events. Thus his instance of an 
analytical judgment is, 'all bodies are extended (or impene- 
trable),' which is ' not an empirical judgment, but stands 
firm a priori ' ; but he would regard this judgment as a 
most important factor in the constitution of knowledge. It 
is not, as an analysis of an empirical conception would be, — 
as with Locke ' all gold is malleable ' would be, — of no use 
in the furtherance of knowledge, a mere account of what is 
understood by a name. On the other hand, ' all bodies are 
heavy' (his instance of a synthetical proposition), is only 
possible because body is conceived as part of the totality of 
experience to which I may add other parts, as I do when I 
VOL. II. q 

226 LOGIC : MILL. 

recognise by observation that bodies are heavy. 1 But the con- 
ception of experience as a whole of mutually determining parts 
is, as it is Kant's great merit to have shown, not of empirical 
origin. It arises, as he awkwardly puts it, from the ' I think 
which accompanies all our representations,' — from the ' unity 
of the self-conscious subject present to all intuitions.' 

69. Now a judgment determined (rendered possible) by such 
a conception is no longer of empirical origin in the sense of 
being a statement of an event, or of a multitude of past events, 
or of an expectation resulting from such ; and the question is 
whether, when the meaning of 'judgments of experience ' has 
been thus altered from what it was with the school of Locke, 
the antithesis either between them and ' a priori ' judgments, 
or between synthetical and analytical, can be maintained in 
the absolute form in which Kant puts it. No doubt Kant is 
right in holding that there are universal and necessary 
judgments which are not analytical in that sense in which, 
according to the Lockeian (experiential) doctrine, all uni- 
versal judgments must be, i.e. as evolving the connotation of 
a name ; and in maintaining the synthetical character both 
of mathematical judgments, and of universal judgments 
about nature, e.g. ' all bodies have weight,' — synthetical, i.e. 
as opposed to the above sense of analytical. The question 
is whether his own doctrine is not fatal to the acceptance of 
the distinction between analytical and synthetical judgments 
in the form in which he puts it, i.e. not as a distinction 
between verbal and real propositions (propositions evolving 
the connotation of a name and propositions stating a matter 
of fact), but as a distinction between two sorts of real pro- 
positions, — those stating 'judgments of experience' (syntheti- 
cal) on the one hand, and those stating ' a priori ' truths 
(analytical) on the other. If a proposition which is to state 
real knowledge about things must state something which 
happens or has happened, one can see (a) that such a 
proposition cannot take universal form without ceasing to 
state real knowledge, and (b) that it is absolutely opposed 
to any statement about a conception in the mind. But if 

1 • Thus it is experience upon which (only contingently, however), as parts 

rests the possibility of the synthesis of of a whole, namely, of experience, 

the predicate of weight with the con- which is itself a synthesis of intuitions, 

ception of body, because both concep- - Critique of Pure itacm (Meiklejohn's 

tions, although the one is not contained Trans.), p. 8. 
in the other, still belong to one another 


' all bodies are heavy,' explained as Kant explains it (without 
any attempt to reduce it to a statement of events), is to be 
our instance of a ' judgment of experience,' (a) is the differ- 
ence between conceptions which merely ' belong to one another,' 
and those of which c one is contained (though confusedly) in 
the other,' such as to justify the assertion of an absolute 
difference between synthetical and analytical propositions ? (b) 
is ' all bodies are heavy ' — a synthetical judgment of experience 
— still so far from being universal and necessary that when 
we find a judgment which indisputably is universal and neces- 
sary, we have no alternative but to suppose it either not syn- 
thetical (merely analytical), or, if synthetical, then not 'of 
experience ' but ' a priori ' ? If we answer (b) in the negative 
we have to alter Kant's distinctive doctrine of the relation 
between mathematical and other knowledge; if (a) in the 
negative, we must reconsider his doctrine as to the impossi- 
bility of extending knowledge beyond the region of possible 
intuitions, the point of this doctrine being that without in- 
tuitions there are no synthetical judgments, and any exten- 
sion of knowledge must consist in synthetical judgments. 

70. In judging body to be extended, according to Kant, 
we do not go beyond the conception of body ; in judging it 
heavy, we do go beyond the conception ; the conception of 
heaviness belongs to, but is not, like that of extension, 
contained in, that of body. But do we not go beyond the 
subject- conception ' body ' in judging it to be extended ? 
The judgment is otiose unless we do. In all real thinking 
such a judgment would represent a process of going beyond 
a given subject to connect it with others, — either to connect 
this body with other bodies as limited by and limiting them, 
or to connect body under the point of view of its being a 
composition with all things else divisible. From Kant's 
point of view it might be said that in one sort of judgment 
we go beyond the subject indeed, but only to connect it with 
other subjects in virtue of properties belonging to thern as 
the result of forms of intuition ; in the other we go beyond 
the subject to connect it with others in virtue of properties 
given in empirical intuition. But since with Kant both the 
existence of objects of empirical intuition, and the connection 
of one object with another in a nature, are rendered possible 
by synthetic intelligence, this distinction reduces itself to 
one of degree in complexity of synthesis. In both cases, in 

Q 2 

228 LOGIC : MILL. 

the actual process of knowing, you go beyond the subject, 
but in one case to connect it with a more, in the other with 
a less, complex result of intellectual synthesis. 

71. To avoid misapprehension, we must bear in mind 
(what Kant would not have disputed) that both propositions 
alike, as representing the mental act of the individual, may 
represent one of mere analysis of an accepted definition. 
* Under the term e body ' we agree to understand the attribute 
of extension, or that of heaviness, as the case may be.' 
There is no doubt that to all of us f body ' has come to 
include heaviness in its connotation just as much as exten- 
sion. What we have to do is to compare 'all bodies are 
extended ' with e all bodies are heavy,' (1) as we may suppose 
these several judgments first arrived at, (2) as they are em- 
ployed in thinking about the world, about matters of fact, 
in arriving at further truth, as distinct from the process of 
reviewing the connotation of general terms, (3) as represent- 
ing truths about the world. Is there any difference of kind 
between them in these respects, such as Kant supposes 
between analytical and synthetical judgments ? As to (1), 
Kant's own doctrine implies that c all bodies are extended ' 
could not be got at without intuition. It='all bodies are 
parts of space,' and this implies the presentation of ' pure 
intuition.' It is not analytical, then, as first arrived at, 
but, according to Kant's doctrine, it is arrived at (a) by a 
synthesis more primitive than that by which 'all bodies are 
heavy ' is arrived at, so primitive, indeed, that without it 
the conception of body could not be formed at all, (b) by a 
synthesis exercised upon a pure form of intuition, as distinct 
from a synthesis exercised, as in the case of the other 
judgment, upon an empirical intuition. This does not mean, 
however, that ' all bodies are heavy ' is a summary of events 
in the way of sensation, according to Kant, any more than 
the other. An intellectual synthesis is necessary to give it. 
6 Weight ' is not a mere feeling, but an < empirical con- 
ception ' resulting from the interpretation of feeling under 
the direction of the ' synthetic principles of understanding ' 
(in particular the principle of 'anticipations of perception'), 
and as predicated of body it implies the conception of the 
connexion of body with the ' whole of possible experience.' 
As arrived at by us, then, it is not true to say that one is 
got at by analysis of a conception, the other not, or that one 


is arrived at without experience, the other by means of it ; one 
independent of sense, the other not. Each presupposes a 
synthesis, and in its universal form results from an analysis 
of the result of such synthesis. The difference is that one 
presupposes a less complex synthesis than the other, and 
that while, in order to arrive at one in its explicit universal 
form, the analysis needs only to be of that result of synthesis 
which we call ' body,' it is a more complex result that must be 
analysed in order to yield the other, in a like explicit universal 
form. So far there is truth in saying that in one of the 
two judgments the predicate is contained in the subject, in 
the other only belongs to it. 

72. 1 ' We cannot,' it may be said, ' conceive body at all 
except as extended ; the conception of it as heavy is gradually 
formed through experience.' If it comes to a question of the 
individual man's power of conception, there is quite as much 
ground for saying that imponderable bodies are inconceivable 
as unextended ones. It is true that extension when analysed 
turns out to be the simpler conception. In Kant's language 
it is the conception of that which is the condition of all 
intuition — all perception and imagination. But (a) in this 
respect how does ' all bodies are extended ' differ from 
'a straight line is the shortest between two points,' which 
is equally based on a necessity of intuition ? (b) why, because 
extension is a simpler, more abstract, attribute than heavi- 
ness, should one be said to be contained in that of body, the 
other merely to belong to it ? If we treat it as a question 
of what any individual understands by ' body,' one attribute 
is as much contained in it as the other. If it is a question 
about body as it is really determined in the universe or for 
a perfect intelligence, gravity is as necessary to the con- 
ception of it as extension. It is only relatively to us, as of 
growing intelligence, that the distinction between the simpler 
attribute as contained in, and the more complex as belonging 
to, a subject, can have any meaning, and for us it is only a 
distinction of degree. In like manner, if the distinction 
between analytical and synthetical judgments is meant to 
be one of truth, as distinct from one of the way in which we 
as individuals apprehend them, for a distinction of kind 

1 [This section is to a great extent between the two kinds of judgment, 

identical in substance with the preced- which are stated but not discussed in 

ing one, but it deals briefly with the the previous section.] 
eecond and third heads of comparison 


should be substituted one of degree between more elementary 
and abstract truths on the one side and more determinate 
and concrete on the other. Universality and necessity no 
more belong to one sort of truths than to the other, though 
about the latter we are more likely to make mistakes. As 
regards analysis and synthesis, the truth is that every 
judgment (a) presupposes a synthesis which is an analysis 
of the confused, (b) is itself an analysis, (c) an analysis 
which in all actual thinking is a step to farther synthesis. 
A synthesis of sensations so as to form an intensive quantity 
is necessary to the conception of weight; a further synthesis 
is necessary to the conception of this as an attribute of body. 
The result of this synthesis is that conception of body which 
is analysed in the judgment ' all bodies are heavy,' of which 
in actual thinking the purpose would always be some 
synthesis, such as the connection of the motion of a particu- 
lar body with other phenomena, as an instance of gravity 
modified by other forces. 

73. But it may be said that, though these objections 
may be valid against the distinction between analytical and 
synthetical judgments, as Kant puts it, the distinction re- 
mains between judgments respecting matters of fact and 
judgments respecting mere ideas. No doubt there is a 
distinction between propositions which merely state what 
the propounder understands by a general name and those 
which apply to reality. The object of the former is either 
(a) the instruction of another, or (b) agreement with an 
opponent as to the sense in which a word is to be used, or 
(c) rhetorical deception, or (d) the clearing up one's own 
thoughts. The real question is (a) whether all judgments 
which do not state events in the way of. feeling (which no 
general proposition can do) are thus merely analytical of 
the meaning of a name, and (b) whether, if we admit judg- 
ments which do not state events in the way of feeling as 
yet relating to matter of fact, the distinction between matter 
of fact (reality) and thought can be maintained. Question 
(a), as we have seen, Locke, with a varying amount of draw- 
back, answered aflirmatively in regard to 'substances' or 
' coexistence in nature,' but admitted judgments respecting 
mere ideas which were yet ' instructive ' and stated real 
truth — were not analytical of the meaning of a name — viz. 
mathematical judgments. This is inconsistent (1) with his 


view of the unoriginativeness of mind, (2) with his view of 
reality as that which happens to us, (3) with his view of 
mathematical truths as representing primary qualities of 
matter. Hume in the ' Treatise of Human Nature ' tries to 
reduce mathematical propositions to statements of matter of 
fact as sensible events, which he admits destroys their uni- 
versal character. In the * Essays ' he reverts to Locke's view. 
He classifies propositions as statements either of matters of 
fact or of relations between ideas, and puts mathematical 
propositions in the latter class without meeting the objections 
to this view which have been stated above with reference to 

74. Mill tries to combine the view of Hume's ' Treatise ■ 
with that of his ' Essays ' without recognising the conse- 
quences of either. He identifies mathematical propositions 
with propositions respecting matters of fact — reduces all 
propositions which do not concern matters of fact to ana- 
lyses of nominal essence — and at the same time ascribes uni- 
versality to propositions concerning matter of fact, which 
Locke and Hume had seen could not belong to them if 
they = statements of events, and could only belong to mathe- 
matical judgments on the supposition that they concerned 
mere ideas. 

In truth, Mill's view of mathematical propositions, as 
well as many other propositions concerning nature, is in- 
compatible with their being statements of events at all : it 
implies that they are statements of conceived relations be- 
tween objects which are not events at all, and can only be 
called phenomena in the loose sense of that term in which 
it stands, not merely for a sensible event, but for any object 
of consciousness whatever. His other universal propositions 
concerning matters of fact, though they relate to events, 
relate to them as determined by an order of nature which 
is not an event or sum of events ; and it is only in virtue of 
determination by such an order that events can become 
subjects of universal propositions. 

75. When matters of fact (or ' phenomena ') have thus 
ceased to be mere ' events in the way of feeling,' — have come 
to be regarded as appearances of an order which, as can be 
shown, can only exist for a thinking as opposed to a merely 
feeling subject, — the question is whether the distinction 
between them and ideas (as = thoughts as such) can be 

232 LOGIC : MILL. 

maintained ; whether for this must not be substituted the 
distinction between them and the meaning of general 
names, which = the conception on the part of the person 
who uses the name as it happens at any time to stand, 
which presumably is untrue and inadequate. In regard to 
propositions professing to concern nature, then, for Locke's 
distinction between those expressing real existence and 
those explanatory of a complex idea, for Kant's between 
synthetical and analytical, we shall adopt another distinction 
of Kant, between those which do and those which do not 
concern objects of possible experience, understanding at the 
same time that thought is necessary to constitute experience. 
In other words, all universal judgments that are to be more 
than nominal, though they can by no means be reduced to 
singular ones, any more than conception can be intuitional- 
ised, must be verifiable by an intuition, which could be 
stated in a singular proposition. 



(Mux, Book I., Chapter VIII.) 

76. No definition is a statement of a present sensible 
event, or of a multitude of such events. Accordingly, 
according to Locke's view of propositions concerning real 
existence, no definition (subject to a reservation in favour of 
mathematical definitions) can concern real existence : it can 
only be an analysis of nominal essence. In like manner, 
Mill, at the beginning of the chapter on definition, calls 
them the ' most important of propositions purely verbal.' 
The natural meaning of this is that definition does not 
relate to matter of fact or reality, that it merely analyses 
a, complex idea as opposed to stating what happens (which 
alone = the real). But, as we have seen, Mill is am- 
biguous in his account of ' verbal propositions.' If every 
proposition is ' merely verbal ' which asserts something of 
a thing c under a name that already presupposes what is 
about to be asserted ' (I. vi. § 5), then every statement by a 
scientific man is for him merely verbal. ' Water is composed 
of O and H in such and such proportions.' In calling any- 
thing ' water,' the scientific man understands himself to 
imply that it is so composed. Such a statement, however, 
by no means falls under Mill's other account of a * verbal 
proposition,' viz. that it is one which ' does not convey 
information.' So with every definition: it specially purports 
to be a statement of the full meaning with which the definer 
uses a certain name ; thus it is necessarily • merely verbal ' 
in the sense that it * asserts something of a thing under a 
name that ' to the definer e presupposes what is about to be 
asserted.' But it does not in consequence ' convey no 

77. The truth is that every general proposition comes 
under the above account of i propositions merely verbal,' and 
a strict -follower of Locke and Hume would have to admit 

234 LOGIC : MILL. 

that no such proposition concerned real existence, or con- 
veyed any information except as to the usage of a name. 
Mill, however, is by no means prepared to admit this : he 
holds that general propositions may represent real existence ; 
that the question whether they are merely verbal or no 
depends on whether the conception, which such a proposi- 
tion in every case unfolds, represents a real union of 
attributes in a thing. This being so, the distinction between 
verbal and real propositions, as he puts it (i.e. in a form 
only suitable to the view which regards every general proposi- 
tion as merely verbal), loses its meaning, and in consequence 
we find him constantly treating propositions as real, not 
verbal, which yet fall under the above description (e.g. ' all 
men are mortal ') . In like manner, when treating of de- 
finition, he is still so far affected by the Lockeian theory 
and his own account of ' verbal propositions ' as to identify 
definition with verbal proposition, but soon comes to write 
of definition in a way that makes this unmeaning. Thus 
(I. viii. § 1), as a formula for expressing the definition of 
c man,' having first adopted c man is a name connoting such 
and such attributes,' he immediately substitutes c man is 
everything which possesses such and such attributes,' thus 
showing that he does not consider the definition to be 
merely an analysis of nominal essence in Locke's sense, but 
a statement of a real coexistence of attributes in a real 
subject. Again, having said (ib.) that a definition is ' the 
sum total of all the essential propositions' (which = propo- 
sitions merely verbal) ' which can be framed with a given 
name for their subject,' he says (ib. § 3) that ' the only ade- 
quate definition of a name is one which declares the facts, 
and the whole of the facts, which the name involves in its 
signification.' Does this mean all the qualities, which those 
who apply the given name to an object understand that 
object to possess (the object being a creature of their 
thoughts) ? Or does it mean that the content of a definition 
should be some group of qualities or phenomena really 
connected with each other and constituting a real object to 
which the defined name is applied ? If it means the former, 
the question being simply one of usage, it will be impossible 
to do more than lay down some meaning in which the given 
word is to be used as between certain persons, or in a certain 
book, or in a certain enactment. To determine the sense in 


which all who use some name happen to use it, unless it be 
a technical term, is clearly impossible. But if this is what 
Mill means by definition, there is no purpose in insisting on the 
difference between 'complete' and 'incomplete' definitions, 
and that the former must state ' the whole of the facts, &c.' 
All that can be reasonably sought for in such a definition is 
a guide to usage, which the ' incomplete ' definition fur- 
nishes. There is no meaning in the requirement that ' all 
the facts ' be covered by the definition, unless it is under- 
stood that the name represents some real thing (however 
this be explained), and that we have to try to make the 
conception which we connect with the name correspond with 
the facts united in the thing. 

78. In regard to ' scientific definitions ' Mill virtually 
admits this (ib. § 4). Their purpose 'is not to expound 
a name, but a classification,' and that not an arbitrary 
classification. As Mill's instances show, their object is to 
state some real relation of properties, which a multitude of 
other such relations really depend upon, or at any rate 
accompany. In other words, their object is to do just what 
the definition of Plato and Aristotle was meant to do, viz. 
to explain the phsenornena of the world as graduated modifi- 
cations of simpler principles; only they do it in a less 
superficial way. They do not find their simpler principle in 
an abstraction from some rough current conception, which 
was what the ancients did. Thus, to take Mill's instance of 
the quest for the definition of heat (which he strangely 
speaks of as an inquiry into the meaning of a word), the 
object is to discover the 'power which causes what our 
senses recognise as heat' (a ' verum genus 9 ), and then to 
differentiate this genus by determining 'under what charac- 
teristics the multitudes of phenomena certainly connected 
with this power ' may 'be embodied as a class ' (i.e. treated 
as modes of this power), ' which characteristics would of 
course be so many differentia for the definition of the power 
itself (ib.). We may almost say that whereas the ancient 
logic, supposing itself to be defining the nature of things, 
really was but analysing the received meaning of general 
names, the modern logic, while insisting that it is only 
explaining the meaning of words, is really engaged in gradu- 
ally defining the nature of things. 

79. The reason why Mill and others (who, unlike Locke, 


admit the possibility of truths at once real and general), are 
afraid of allowing that definition can ever be of the ' nature 
of things,' is that they think this implies the 'reality of 
essences,' i.e. that there are real things corresponding to 
the signification of general names. For these they would 
substitute real uniformities in the relations of phenomena, 
to which they would rightly hold that the meanings of 
general names seldom correspond. This is a very good 
reason for refusing to admit that those definitions which 
merely analyse what is ordinarily understood by general 
names are at the same time explanations of the nature of 
things ; but not a reason for denying that either scientific 
definitions, which purport to state some most c general 
uniformity of phenomena ' as modified by particular condi- 
tions, or mathematical definitions, relate to real existence. 

80. Undoubtedly every definition is an analysis or 
explanation of the meaning of a name (of a conception 
represented by a general name), though in some cases not of 
a name previously in vogue, but of one introduced to 
represent the conception which the definition states. This, 
however, is saying very little. Everything depends on the 
nature of the conception, and the sense in which the definition 
professes to explain it. It may be (a) a conception current 
among men, of which the definition undertakes to explain 
the ordinary content, what it involves for most men who 
entertain the conception. Only such a definition can fitly 
be said to be ' merely of the meaning of a name.' (b) It 
may be a conception which the definer does not find in 
vogue, but undertakes to constitute. Such are (1) defini- 
tions contained in law (the definition, say, of manslaughter, 
or of a. public elementary school), and those which the 
judge and jurist derive from these : (2) definitions of duties 
by the moralist, so far as he undertakes to do more than 
expound common sense : (3) mathematical definitions ; the 
definition of a circle represents the mental act of construct- 
ing it. (c) It may be a conception representing the definer's 
discovery of facts of nature, or his analysis of metaphysical 
conceptions, — those conceptions which do not properly result 
from experience, but regulate its formation and the growth 
of usages, institutions, and practical ideas among men, e.g. 
cause, substance, right. 

81. In regard to (a) and (c) completeness of definition is 


impossible. In the case of (a) there is no object in seeking 
it ; what is wanted is either a guide to usage, or security 
for consistency in the use of a term by writers or disputants. 
In the case of (c) there is a constant progress towards 
completeness of definition, but definition is the end of the 
scientific process, not the beginning ; i.e. it is a constant 
effort to reach ultimate principles — force, configuration of 
particles, primitive cell, or tissue — of which all phenomena 
may be exhibited as a graduated modification; in other 
words, to exhibit phenomena as successive differentiations 
of a genus. If by definition is meant the analysis of a given 
conception represented by a word, the scientific process has 
nothing to do with it; but if it means the process of 
reaching such a conception as above described — the true 
genus duly differentiated — then all science consists in or is 
subsidiary to definition. As to metaphysical conceptions, 
their full definition is only supplied by the scientific or 
practical experience which they regulate. Such definitions 
as (b, 1) are (1) liable to alteration by new law, (2) only 
attained so far as law takes the form of explicit and consist- 
ent general enactment. In fact, a great part of the business 
of the lawyer consists in trying to make definitions which 
shall be complete in the sense of covering all usages of a 
term in common law (and even in inconsistent statutes). 
(b, 2) are necessarily incomplete : the genus of a duty may 
be stated, but its differentiation depends on circumstances 
which cannot be determined a priori. Take e.g. the duty 
of truth-speaking. Granted the duty to convey to everyone 
with whom one has to do the most correct notion of what 
one thinks, and the fullest information about all that concerns 
oneself and him, possible, or of which he is capable, or of 
which circumstances allow; all practical questions turn on 
the right interpretation of these qualifications. 



(Mill, Book II., Chapters V., VI.) 

82. There remain (b, 3). 1 The question as to the 
nature of mathematical definitions and axioms, and of 
reasoning apparently founded on them, turns on what we 
consider space itself to be ; space, and not merely our ideas 
of space. It is characteristic of Mill, Spencer, and the rest, 
that they treat the question at issue between them and 
Kantists as if it concerned merely the origin of our beliefs 
about space. Space itself and its properties they take for 
granted as something of which we have sensitive experience. 
This granted, it is easy to show that our beliefs about 
space are derived from such experience. All that they 
suppose has to be accounted for is the strength of convic- 
tion attaching to such beliefs, of which Spencer is considered 
to have found the final explanation in the ' discovery ' that 
this conviction is not derived merely from each individual's 
experience, but from the experience of endless generations, 
of which the result in the way of strengthened belief is 
transmitted from one generation to another. That this 
'discovery' should be supposed to have any bearing on 
the real question at issue, shows how entirely this question 
is misapprehended. No explanation of the readiness and 
strength of conviction with which the individual accepts 
certain beliefs about space amounts to an explanation of 
what space itself is. 

83. Is space a sensation? Berkeley would have held 
this if he could ; but so soon as he has used language which 
implies it (which implies the identification of visible 
extension with colour), he substitutes for it language which 
implies at least a multiplicity of sensations. 2 Is it then a 
succession of sensations ? This is the only logical alternative 
to those who, like Hume, accepting Berkeley's reduction of 

1 [Above, section 80.] 

2 See General Introduction to Hume, sees. 177-178, vol. i. pp. 144-145. 


the sensible thing to sensations, also reduce the mind to 
a flux of feelings. Space, however, is supposed to be an 
aggregate of coexisting parts, and a succession of sensations 
cannot constitute such an aggregate. Thus Hume has to 
regard it, not as a sensation, but a compound of feelings, 
without explaining how such a composition is possible. He 
makes space (extension) a compound of which the ultimate 
parts are not extended (spaces), but are sensations of colour 
or hardness. He avoids, however, the difficulties that arise 
from representing parts of a space as not themselves spaces 
but feelings, by substituting for ' sensations (impressions) of 
colour,' &c. ' points or corpuscles endowed with colour and 
solidity.' Thus he really assumes space in his account of 
the impressions, of which space is to be explained as the 
compound. His doctrine, as he admits, renders the universal 
propositions of geometry not only untrue, but unmeaning. 
There is nothing in reality or in the mind corresponding to 
the right lines of which the mathematician asserts that they 
can never have a common segment ; (the only right lines 
which exist either as impressions or ideas, in reality or in 
the mind, may have a common segment ;) nothing to the 
isosceles triangles of which the mathematician says that the 
angles at their bases are always equal, &c. &C. 1 

84. Kant's doctrine of space and time is quite compatible 
with the admission that the abstract ideas of space and time 
have only been gradually attained by the human mind. It 
means that space and time are relations, under one or other 
or both of which all sensible objects are presented to us, but 
which are neither sensations, nor sensible objects, nor results 
of abstraction and generalisation from such objects (since 
they are the conditions of the earliest perception), but are 
constituted by the mind in the act by which mere sensation 
becomes ' intuition ' (perception) . Kant expresses this some- 
times by saying that they are 6 forms ' added by the mind to 
' matter ' given in sensation. The ' priority ' thus claimed 
for space and time is not a priority of the abstract ideas of 
space and time, but a priority of space and time, as relations 
constituted by the mind, to the sensitive experience which 
they determine, and which through them becomes an in- 
tuition of objects. £ Priority ' is an unfortunate term, be- 
cause it suggests antecedence in time; but all that Kant 

1 lb. sees. 274-275; vol. i. pp. 231-233. 


meant by it was that space and time were not conceptions 
resulting from experience, but conditions given by the mind, 
under which sensation becomes an experience of definite 

This view at any rate does not exclude the c historical ? 
view that the detachment of these relations from other con- 
ditions of reality — a detachment by which the abstract ideas 
of space and time are formed — comes late in the progress of 
the human mind. Rightly developed, it is the true safeguard 
against 'interposing the fiction of time between ourselves 
and reality.' If space and time are relations constituted by 
the mind in the act of intuition, they are clearly not con- 
ditions under which the mind itself exists, nor conditions of 
any reality other than a perception. The objects of our 
knowledge are relations, and, according to Kant's own show- 
ing, no relations are in space and time. Only sensible objects 
are in space and time, and relations are not sensible objects. 

Kant's great mistake lay in holding that the only objects 
of knowledge were objects of 'possible perception,' from 
which it followed, since space and time were conditions of 
perception, that nothing could be known except under these 
conditions. He was strong, however, against admitting that 
nothing could be thought except under these. It is thus 
incorrect to represent him as having held space and time 
to be in any way, direct or indirect, conditions of thought. 

85. With Kant space is the " form of intuition ' of a cer- 
tain sort, viz. outward, for which he occasionally substitutes 
' form of outer sense ' ; the ' form ' in virtue of which objects 
are perceived or imagined as external to each other and to 
our organism. Upon this it may be asked, (a) What is meant 
by ' form ' ? (b) When you define space as the form of outer 
sense, or the form in virtue of which objects are perceived 
or imagined as external to each other and to our organism, 
are you not assuming the thing to be defined? are you not 
in effect saying that space is the form in virtue of which we 
perceive things in space ? Where Kant writes s form,' we 
may generally put ' relation.' With Locke all ' relations ' 
are creations (fictions) of the mind (understanding). He so 
reckons them because no single feeling (simple idea), nor any 
number of such feelings, except as combined by a subject 
other than any of the feelings, can constitute a relation, and 
only such feelings are given to the mind (and thus only they 


with Locke are real) . In Kant's language what is given to 
the mind in feeling is the ' matter ' of sensibility ; the rela- 
tions by which this ' matter ' is determined, i.e. which arise 
from the holding together of feelings by the one mind present 
equally to them all, constitute the ' form.' These, according 
to the same mode of speech by which the feelings are said 
to be given to the mind, are said to be added by the mind. 
The great mistake to guard against (and of which Kant by 
no means keeps clear) is that of supposing the ' matter ' of 
sensibility to be really anything apart from ' form,' — objects 
first to be felt and then relations added. This does not mean 
that feeling cannot take place without thinking, but that for 
a merely feeling consciousness there is nothing of which reality 
can be predicated, no real objects ; though a feeling con- 
sciousness itself becomes a real object for a thinking con- 
sciousness. Thus ' space is the form of outer sense '= space 
is the relation by means of which the mind constitutes the 
outwardness of sensible objeets. 

86. Now it is quite true that, since outwardness = exist- 
ence in space, this is not a proper definition. It does not 
resolve a complex conception into genus and differentiae, into 
a simpler conception qualified in a particular way. Space 
being an absolutely primary and simple relation cannot be 
thus analysed. A relation of another kind might be explained 
by a statement, not implying the relation in question, of what 
the objects are between which the relation obtains ; but this 
cannot be done in the case of space, because, owing to the 
primariness of the relation, objects determined by it need 
have no nature except what they receive from the relation. 
What the above account of space does, is to bring out (a) 
that space is properly a relation, and as such constituted by 
the mind, and (b) by use of the term ' outward ' (which, 
though it strictly = * in space,' is, for reasons which will 
appear, less liable than • space ' to be taken to signify a 
'thing in itself), what the qualification of the object arising 
from this relation precisely is. State the qualification as 
outwardness, and it becomes apparent, (1) that it is ultimate 
— cannot be analysed into anything simpler ; (2) that it is not 
a condition of all objects (not of object qua object) ; that, 
e.g. though a qualification resulting from a relation, it is not 
one by which any relation can itself be qualified, since no re- 
lation is outside another ; that thus (3), though a qualification 



coextensive with the conditions of perception and imagination 
(since these we take to be relative to objects outside each 
other), it is not a qualification of objects as thought of. (2) 
and (3) are propositions implied in Kant's doctrine that 
space is not a 'thing in itself or a quality of 'things in 
themselves ' ; a doctrine which has been misunderstood 
because Kant himself puts as the equivalent for it the state- 
ment that ' space is only subjective.' No one is clearer than 
Kant that space is a condition of all possible objects of ex- 
perience. But what is true (and what on the whole Kant 
meant) is (a) that space is not a qualification of things as 
apart from mind or intelligence ; (in this regard Kant's fault 
consisted in often writing as if things might so exist ; this is 
one of the senses in which he uses ' things in themselves ') ; 
(b) that space is not a qualification of objects of thought, as 
such (Kant would say that it was a qualification of all objects 
of knowledge, because only what can possibly be perceived can 
be known), nor, in particular, (c) of the mind or ego, from 
whose synthetic action space results. 1 

87. So far Kant's doctrine seems irrefragable. It is the 
logical result of the failure of Hume's attempt to treat space 
as an aggregate of feelings. The rejoinder will be, 'It is neither 
an aggregate of feelings, nor a relation between felt objects 
constituted by thought, but an attribute of that matter which 
causes our feelings, and is revealed through them.' We are 
quite agreed that it is an attribute of matter, but what and 
whence is matter ? ' Unknown ; we only know that it pro- 
duces effects in the way of feeling.' But, if it is unknown, 
you are talking nonsense in saying that it produces these 
effects, and in saying that space is an attribute of it you have 
in effect said nothing. We say that this mystery about 
matter is a mystery of your own making. Matter is a con- 
geries of relations constituted by thought ; resulting from the 
presence of thought (a thinking subject) to feelings, of which 
relations the simplest is space. ' How can that be, when 
thought is a result of matter P ' But you have just said that 
matter is the unknown, as, in abstraction from all relations 
which can be shown to be creations of thought out of feelings, 
it undoubtedly is. Which, then, is more rational ? To try 
to explain thought as a result of matter which cannot be 

1 H. Spencer says that, according to Kant, ' space and time are conditions of 
the ego? 


known, or matter as a result of thought which is known ? 
Your objection to the latter course is really due to the fact 
that our individual intelligence is gradually developed, and 
presupposes a world, of which matter is one condition among 
others, as determining its development. This, however, is 
no ground for supposing this world to be one in which 
thought is not (which is what it is apart from thought), which 
in fact is to make it simply the negation of intelligibility, but 
only for holding that there is a complete subject-object, a 
complete intelligence including its own object, on which our 
intelligence depends. 

88. The possibility of space, then, presupposes that 
presence of thought to, and action upon, sense, which con- 
stitutes sensible objects other than, but determined by 
relation to, each other ; and space consists in a particular 
mode of this relation, this otherness ; a particular mode of 
the relation which does not obtain between all, even sensible, 
objects, not, e.g. between sounds. (One may represent to 
oneself the distinction between two sounds as an interval of 
space, but this is Kara (jusTCKpopdv ; one sound is not outside 
the other, as one visible object is outside another.) Then 
this relation between objects (relation of mutual limitation) is 
itself considered as an object ; or, to put it otherwise, the 
objects which it qualifies are considered as having no other 
qualification than that which they derive from the qualifi- 
cation. The possibility of so considering them arises from 
the primariness of the relation. Being the condition of all 
perceivable objects, it can be supposed present without any 
other conditions, but none other without it. We thus get 
mere spaces, and space as the aggregate of such. 

89. Of space in this sense it is true enough to say that 
it is an abstraction, in the sense that it is an object consti- 
tuted by separation of one relation by which real objects 
are determined from all other relations. But we cannot 
abstract what is not there to be abstracted. Space, as a 
relation of the kind described, is not an abstraction, but 
abstraction of this relation from all others yields the object 
called pure space. It is a mistake, however, to suppose 
that abstract space (or our abstract idea of space) is an 
abstraction from sensitive experience, for that which is thus 
abstracted from experience (separated from other conditions 
of experience) is only in experience just so far as experience 

B 2 

244 LOGIC : MILL. 

is not merely sensitive. When Kant speaks of space as a 
condition of experience, he means space as a relation, not 
space as an abstract object. He represented this conversion 
of space as a relation into space as an object, by saying that 
space was not merely a form of intuition, but itself an intuition. 

90. An object considered simply under the relation of 
externality (mutual limitation) to other objects all about it 
is space in three dimensions : the relation of two such spaces 
to each other constitutes surface ; of two surfaces to each 
other, the line ; of two lines to each other, the point. Then 
arises a science of which the materials or objects consist 
simply in the various forms of this abstraction of the limit ; 
a science which, assuming the mere point, mere line, mere 
surface, proceeds by putting them together to find new 
figures with new properties. 

A point is the simplest or most abstract form of the 
limit — the relation between spaces not considered as itself a 
space. (Thus considered, it has no parts. Considered, not as 
the mere limit between lines, but as a component part of a line, 
it must itself, like every quantum, have parts.) A straight 
line, as ( lying evenly between its extreme points ' and 
having surfaces on each side of it exactly alike (not one 
convex, the other concave), is the simplest or most abstract 
form of limit between surfaces, in the construction of which 
nothing is assumed but points. A plane is the simplest 
form of limit between solids. All sorts of angles may then 
be formed by combination of such lines. All sorts of lines 
not straight, and not resoluble into lines at an angle with 
each other (curves), may be drawn, varying according to the 
mutual relations of straight lines drawn from points through 
which the line passes to a given point. In this nothing is 
taken for granted but points, straight lines, and the possi- 
bility of drawing a curved line. All sorts of plane figures 
may be formed by combining straight lines under various 
relations to each other. 

All varieties of figure are varieties of mere boundary or 
limit. The curved line differs from the straight in virtue of 
the different way in which it bounds the surfaces between 
which it is the boundary. The differences of angles consist 
in different relations of boundaries, of lines having no 
properties but as boundaries. All these figures are formed 
out of a material (constituents) purely ideal, or having no 


sensible attributes, since it is obtained by detachment and 
substantiation of a relation constituted by the action of 
thought upon sense, — detachment of that element in per- 
ception which is not sensation (the non-sensuous element in 

91. Ideal figures thus constructed (in the definitions), 
geometrical science inquires into their properties, i.e. into 
their relations in the way of quantity to each other. This 
implies the assumption of the idea of quantity — of wholes 
related to each other as greater, equal, and less in virtue 
of the sums of their parts, increasing with addition and 
diminishing with subtraction of parts. The statement of 
this idea forms the first axioms of Euclid. The 'proper 
axioms ' seem to me inseparable from, or given in, the 
performance of the mental act represented by the definitions 
of the objects to which they relate. One cannot draw a 
straight line according to the definition without seeing that 
there can only be one such. Put the definition in the form 
*■ a straight line is the shortest between two points,' and the 
axiom (which =< there is only one shortest way') is explicitly 
involved in the definition. In like manner, since there is 
only one way of being equal, though infinite ways of being 
greater or less (the relation of equality is always one, while 
the relations of greater and less are infinitely various), the 
angles formed as the definition states right angles to be, 
viz. by one line falling on another in such a way as to make 
the adjacent angles equal, must always be equal to each 
other. (Given the definition, and the conception of equality 
which it involves, the axiom is involved too.) The third 
axiom, as Whewell states it, 1 seems the direct negative 
consequence of the definition of parallel straight lines. The 
definition is, ' Parallel straight lines are such as are in the 
same plane, and which, being produced ever so far both 
ways, do not meet.' The axiom, as Whewell says it may be 
stated, is, ' Two straight lines which cut one another are 
not both of them parallel to a third straight line.' This is 
involved in the definition thus. The immediate consequence 
of the definition is that ' lines which anywhere meet are not 
parallel ' ; therefore lines which anywhere cut each other are 
not parallel to each other ; therefore not to a third straight 

! History of Scientific Ideas, b. i., p. 100. 


92. Geometry rests on definitions, i.e. on certain mental 
acts, the performance of which the definitions represent. 
These definitions differ (a) from analyses of the connotation 
of general names, (b) from those definitions of the sciences 
-which represent a certain stage of discovery, in being fixed 
and final. Kant was quite right in saying that the judg- 
ment c a straight line is the shortest way between two 
points ' is synthetic and a priori, in the sense thafc it is not 
analytical of a conception previously formed ( c a nominal 
essence '), and that it does not mean ' has always been found 
to be so, and is expected always to be found so.' It repre- 
sents an act of mental construction, the result of which no 
experience can modify (thus is true a priori) ; but, inasmuch 
as it presupposes detachment from concrete objects of the 
relation of limitation, it may from that point of view be 
reckoned analytical. 

When Kant says 1 that ' intuition must lend its aid ' in 
order to the possibility of the judgment ' a straight line is 
the shortest way between two points,' he means that we 
must mentally draw the line ; present in intuition some- 
thing corresponding to the conception of a straight line as 
a quantum. 2 ' Intuition,' according to him, is necessary to 
all synthetical judgments, all judgments that extend know- 
ledge. Having a conception of gold, for instance, through 
an empirical intuition corresponding to it I arrive at the 
synthetic judgment that gold is soluble in aqua regia. 
But because the intuition in this case is ' empirical,' not 
' pure ' — because it occurs under conditions which I cannot 
command or completely know — a universal and necessary 
judgment cannot be founded on it. On the other hand, 
when I draw a line, the intuition is my own making, and is 
made in virtue of the condition of sensibility which deter- 
mines all possible objects of experience. The judgment 
which it yields, therefore, is valid for all possible objects of 

93. All that Mill's objection to the doctrine that geo- 
metry rests on definitions really amounts to is this, that 
it does not depend on the use of any particular name. It 
would be just the same if, e.g. what we call a circle, in the 
strict mathematical sense, were called something else. This 

1 Krit. d. r. V., Introduction, v. 

2 lb. pp. 478 and 481-2 (ed. Hartenstein ), pp. 435 and 439, Tr. 


is obvious enough. The science depends on the ideal draw- 
ing of the figures. But why object to call such ideal 
drawing definition ? There is no reason except the notion 
that definition merely has to do with names, a notion which 
he openly departs from in regard to scientific definitions, 
and which he renders unmeaning by the admission (I. viii. 
§ 5) that every definition implies the ' postulate ' either of 
an idea or a thing corresponding to the definition. 

94. Given definitions and axioms, propositions about 
the properties (quantitative relations) of the figures are ob- 
tained by constructions which enable us to compare imme- 
diately as wholes, or remainders of equals or unequals, lines 
and angles which we cannot compare directly. Character- 
istic of the process is that, a singular proposition having 
been arrived at, — as the result, say, of a mediate comparison 
which a certain construction has enabled us to make of this 
square on the hypotenuse with these squares on the contain- 
ing sides, it is forthwith converted into a general proposition, 
so that in other demonstrations every figure which can be 
represented as the square on the hypotenuse is taken to be 
equal to any two other figures which can be represented as 
squares on the containing sides. 

Observe the way in which the fifth proposition of Euclid 
is arrived at. The proposition that triangles having two 
sides and the contained angle equal are equal altogether, is 
obtained by simple superposition of one such triangle on 
another, under the condition that the straight lines forming 
the bases of the two triangles are such as have been con- 
structed according to the definition of straight lines (each is 
the shortest way between two coincident points). Then by 
means of a certain construction the isosceles triangle is 
included in two triangles, which must, according to the 
previous proposition, be equal, because they are formed by 
equal additions to its equal sides, and its contained angle 
is common to both. Thus included, the angles at its base 
become remainders of equal angles, from which angles, 
shown by means of the same construction and application of 
proposition 4 to be themselves equal, have been subtracted. 
In this process observe that though the 'proof may be 
thrown into a series of syllogisms, in each of which either 
proposition 4 or one of the axioms of quantity is a major 
premiss, under which a particular instance is brought in the 


minor, yefc these syllogisms would not enable us to take one 
step towards the proposition demonstrated but for the con- 
struction, in which this single isosceles is included in these 
two several triangles. ■ We are certain, however, that this 
isosceles can be taken to stand for all, these triangles for all 
triangles in which an isosceles can be included in the same 
way/ True, but we are only entitled to this certainty be- 
cause these triangles, being made by ourselves out of elements 
obtained by abstraction of that which is the condition of all 
possible outer sense, are not liable to modification by other 
conditions. A proposition about them true once is true 
always, because it is not liable to modification by new 

95. If the proposition that the square on this hypo- 
tenuse = the squares on these containing sides, which is 
what is alone proved by the demonstration, represented a 
sensible event now happening, and if the propositions involved 
in the demonstration represented sensible events of which 
we only know that they have happened, the certainty could 
not go beyond the ' bare instance.' We should be no more 
entitled to take it as universally true than on the strength 
of an occurrence today to assume that the like will always 
occur again. If, on the other hand, the propositions on 
which the demonstration rests were analyses of what we 
agree to understand by general names, the science could 
never take a step in advance — never could reach an 'instruc- 
tive proposition.' Nothing more could be stated by the 
47th proposition of Euclid than was already involved in the 
definitions. The progress of the science depends on the 
constant mental construction of new single .figures, each of 
which yields a new singular proposition, and each such sin- 
gular proposition, since the figure to which it relates is sub- 
ject to no conditions other than those which the mind gives 
in the act of construction, is equivalent to a universal. If 
one precisely knew all the conditions of a natural event, one 
would be entitled to say not merely (in singular form), this 
event now follows on these, but, universally, such an event 
always follows on these conditions. Where the ' method of 
difference' can be perfectly applied, this universal conclusion 
from a single instance is drawn. Now the figures between 
which the geometrician demonstrates equality, as well as the 
figures (lines and angles) by intervention of which the de- 


monstration is effected, having no properties but what the 
geometrician has either given them in his definitions, or 
demonstrated of them without employing any but denned 
figures, are exhaustively known. Therefore the proposition 
stating the equality is not liable to modification through 
discovery of previously unknown conditions, as a proposition 
stating a relation between a natural event and its conditions 
is, and it can at once take universal form. 

96. We can now see what is to be understood by 
6 mathematical necessity.' It arises from the primariness 
and simplicity of the mental acts constituting those rela- 
tions, by detachment of which from all other conditions of 
the objects qualified by them the mathematician obtains his 
materials. These are mere ' otherness,' and ' otherness ' as out- 
wardness or space. Detachment of the relation of ' other- 
ness ' yields objects having no quality except what is given 
by this relation, i.e. mere units, combination of which is the 
act of counting, and results in the notion of a whole made 
up of homogeneous parts. Such 'necessary truths ' as 2 + 2 = 4 
represent the mere act of counting. The first nine axioms 
in Euclid represent the analysis of the conception of a whole. 

The definitions of geometry, as we have seen, represent 
mental acts, something which we make, not which we find. 
These are not necessary in the sense that we cannot help 
performing them. Most men never perform them, never 
present to themselves a straight line or circle according to 
the definitions. They are necessary in this sense, viz. that 
mutual limitation is the condition of all perceivable objects, 
that it is by detachment of this necessary condition that we 
are able to present to ourselves mere limit in the form of 
surface, line, and point, and that all definitions (or construc- 
tions which they represent) are formed out of the material 
yielded by abstraction of this necessary condition of expe- 
rience. (The axiom about two straight lines is another way 
of putting the definition.) Other geometrical propositions 
are necessary because the objects to which they relate are 
subject to no conditions but those stated in the definitions, 
and thus what has been ascertained of such an object in a 
single instance may be taken for granted without possibility 
of error in all subsequent demonstrations. 

The distinction, then, of the * necessity ' of mathematical 
truths from the ' contingency ' of truths about nature, if it 


is to hold at all, is not to be understood as if it were only in 
mathematics, and not in natural science, that what is once 
true must be always true, or as if natural laws were liable to 
change, mathematical laws not. The true distinction is 
between what is fully true and what is partially true. What 
is fully true once is fully true always, of a natural pheno- 
menon no less than of a geometrical figure ; but any propo- 
sition about a natural phenomenon is true of it only under 
conditions of which we do not know all, while a proposition 
about a geometrical figure, if true at all, is true of it under 
conditions which we completely know. 

97. The doctrine thus stated is not to be confused with 
the doctrine that the propositions of geometry are only 
hypothetically true, i.e. true on the assumption that there 
are objects corresponding to our definitions, which really 
there are not ; that, e.g. there are really no straight lines of 
which it is true that they can nowhere meet, no curved line 
at every point equidistant from a central point, but only 
lines which, as far as the eye can follow them, have so little 
tendency to meet that we can assume them never to do so 
without serious discrepancy between reasoning founded on 
this assumption and the reality, and so with the circle. 
This doctrine results from the notion that the real = the 
sensible. As there are no visible lines of which we can be 
sure that with indefinite extension of our powers of vision 
we should nowhere see them meet, it is at least an assump- 
tion that there are such. But the source of this view is one 
which logically excludes geometrical propositions altogether, 
no less as hypothetically than as really true, the view, viz. 
that whatever exists really or in the mind must be either 
a sensation stronger or fainter, or a collection of sensations, 
or a relation between sensations arising out of the nature of 
the sensations related, or a ' propensity.' Of none of these 
can geometrical propositions be considered true by any 
amount of hypothesis. It is not that they are partly untrue, 
but that they are wholly unmeaning in regard to them. The 
reason why people do not see this is that they allow them- 
selves to talk of ' sensible objects ' without asking them- 
selves what exactly they mean by this. If they did, they 
would see that so far as geometrical propositions relate to 
sensible objects at all, it is in virtue of that in the sensible 
object which is not properly seen or felt, is not a sensation 


of sight or touch, at all. It is impossible to hold that real 
lines are sensations, stronger or more faint ; but people con- 
tinue to suppose them sensible, so that a line is what I see 
or feel it to be, and what I see it to be does not precisely 
correspond to the definition. The answer is, that what I 
properly see it to be has no relation whatever to the definition ; 
it is not a question of more or less precise correspondence, 
but of any correspondence. What I see is colour (or light), 
but line is not a colour nor a relation between colours. To 
make sense you have to say that it is a relation between 
coloured surfaces, but it is not as coloured that the surfaces 
are thus related, for the colours may be changed in any way 
you please, while that relation between the surfaces which 
constitutes the line remains the same. It is only by taking 
' sight ' to express an act of intellectual combination exer- 
cised on materials which are given by sense, but which in 
virtue of this act become quite different from what they are 
merely as feelings, that we can be said to see bodies. The 
idea of the line is obtained by the detachment of a purely 
intellectual element, a mere relation, involved in this ' sight ' 
of bodies. This intellectual element, this mere relation, is 
the real line ; and in geometry nothing is assumed to be true 
about it which is not really true. The notion to the con- 
trary arises from supposing that the real line is a black 
stroke drawn on paper, or some boundary which I see and 
feel between objects, of which I have reason to think that 
with a sufficient magnifying glass I should see that it had 
breadth as well as length, that it was not perfectly straight, 
&c. The truth is the other way. The real line is the line 
made by thought and corresponding to the definition ; the 
seen line is only a line at all so far as the thought of this 
line is superinduced on the sensation of colour. 

In one sense, indeed, geometry is hypothetical; not as 
supposing lines to be straight when really they are not, or 
points through which a circumference passes to be equi- 
distant from the centre when really they are not, but as 
supposing the condition of mutual limitation to be apart 
from other conditions. The limitation is real, and the 
definitions really represent various forms of it, but it is not 
really the sole condition of anything. There are no mere 
spaces as geometry supposes, but other conditions present 
along with that of space do not alter it. 

252 LOGIC : MILL. 


98. Time is the relation in virtue of which an object is 
qualified as having other objects severally before and after 
it ; as beginning upon the cessation of something else, and 
ceasing upon the beginning of something else ; as an event. 
6 This,' it will be said, ' is a bad definition, because in the 
terms ' before, 5 c after,' &c. you assume that which is to be 
defined.' The answer is the same as in regard to space. 1 
Time being a primary relation, its nature cannot be stated 
except in a way which implies that objects, between which 
the relation obtains, are already qualified by it, since they 
need have no other qualification save that which this relation 

It is a relation, then, which implies (1) a manifold, or 
I otherness,' (2) a vanishing manifold, of which (3) the 
vanishing moments are determined by a subject equally 
present to them all, and qualifying one by relation to the 
others. In (2) lies its distinction from space. (1) and (3) 
form the common element which renders objects related in 
the way of time equally numerable, equally capable of being 
treated as quanta, with objects related in the way of space. 

99. Kant calls time, in distinction from space, the form 
of e inner ' sense. This seems a mistake. ' Inner ' and 
6 outer ' are correlative terms, each implying determination 
by the relation of space. They are merely two sides of the 
same boundary : cross it, and to anything now in the ' outer ' 
what was before the ' inner ' becomes e outer.' ' Inner sense ' 
[as ordinarily understood] has derived its meaning from the 
notion that the soul is inside the body, and that one sort 
of consciousness is produced by or through the body, another 
sort produced by the soul for itself. The latter is accordingly 
spoken of as an ' inner sense,' the former as ' outer.' This 
is the ground of Locke's distinction between 6 ideas of sen- 

1 [Above, sec. 86.] 

TIME. 253 

sation ' and l ideas of reflection.' But this distinction is (a) 
untenable in itself, and (6) cannot be made to answer to 
the distinction between objects conditioned by the relation 
of space, and those conditioned by the relation of time, 
(a). Because that to which body is ' external ' is qualified by 
this relation, and is itself from another point of view external 
to that which is external to it. The figure of the soul as 
enclosed within the body, like a box within a box, is the crudest 
possible ; but if it be admitted, to each part of the outer box 
the inner box is external. The only externality is of body to 
body (or of space to space), not of body to mind; while it 
is for the mind that the relation of body to body exists. 
Again, the distinction between perception as produced from 
without, and memory or imagination as produced from 
within, is untenable. A modification of the nervous organism 
is involved in both cases. If it is on account of such 
modification that perception is ascribed to outer sense, 
memory and imagination must be so too. ' No,' it may be 
said, ' but it is owing to the difference between the exciting 
cause of the nervous modification in the two cases that per- 
ception is reckoned c outer,' memory and imagination ' inner * 
sense ; in the one case undulating particles of sether come in 
contact with the extremities of the optic nerve, in the other 
not.' It is not, however, such contact that is the object of 
perception, and besides, though there is a difference between 
the exciting causes in the two cases, it can only be rightly 
described as a difference between an inner and an outer 
cause on the supposition that the exciting cause in the case 
of memory and imagination is mind, and that mind is inside 
body. Neither supposition will hold. It (probably) is not 
mind that excites the nervous modification involved in the 
act of memory or imagination, but some prior modification or 
action of the brain. Without the action of mind in the 
proper sense (the thinking subject), it is true, no object 
would be imagined or remembered on the occasion of such 
modification; but no more without such action would an 
object be perceived on the occasion of an irritation of the 
optic nerve. 

If we still try to make out that the distinction between 
outer and inner sense lies in the difference of the exciting 
cause, whether the difference can properly be called one 
between ' outer ' and ' inner ' or no, we must ask ourselves 


whether or no the distinction intended is one between that 
of which the subject is conscious when the presentation is 
ascribed to ' inner sense ' and that of which he is conscious 
when it is ascribed to ' outer ' : if it is, the difference in the 
physical exciting cause of consciousness will not correspond 
to it, for that of which the subject is conscious in any case is 
not the physical exciting cause of his consciousness. 

100. (b). ' But/ it may be said, ' the perceived object (which 
is quite different from the cause of sensation) is external to 
other bodies and to my body, and for that reason perception 
is referred to outer sense, memory and imagination not.' 
Nay, objects remembered as having been perceived, or 
imagined as possibly perceivable, are external in the same 
sense as perceived objects. They are presented as in space, 
outside other objects. 1 On the other hand, there are ' modes 
of consciousness,' which as implying an external exciting 
cause (of which ' we cannot make one for ourselves ' ) are on 
the same footing as perceptions, and thus should be subject 
to the conditions of space, but which are not in space, e.g. 
sounds and pains. 2 The relation of one sound to another is 
not a relation in space. Sounds, no doubt, are habitually 
referred to objects in space, but the objects in most cases of 
such reference are remembered or imagined, and thus, according 
to Locke's distinction, are objects of inner sense. It must be 
sounds simply as heard with the consciousness of their rela- 
tion (as when we are listening to a piece of music) that are 
objects of outer sense according to Locke's distinction, and as 
such they are not subject to the form of outer sense, i.e. space, 
according to Kant's distinction. 

101. So far, then, we have shown that the ordinary dis- 
tinction between outer and inner sense is untenable, and does 
not match Kant's. If now we adopt Kant's basis of distinc- 
tion, which will require us to ascribe most objects of memory 
and imagination to outer sense, and many of Locke's ' ideas of 
sensation ' to inner sense, the objection still remains (a) that 
the terms ' outer ' and ' inner ' are alike applicable only to ob- 
jects in space, and alike inapplicable to a function of thought, 

1 If you explain the externality of touch with the outward as a feeling of 

the perceived object as the inferred colour. 

possibility of touching it, then the per- 2 These would belong to ' outer 

ceived object is external in a way in sense,' according to Locke's meaning 

which an object imagined or remembered (Essay, II. i. 14; II. ix. 6. Cf. General 

is not; but the weakness of this ex- Introduction to Hume, vol. i. p. 80, note), 

planation is that there is just as much not according to Kant's, 
difficulty about identifying a feeling of 

TIME. 255 

as much to that which constitutes the relation of space as to 
that which constitutes the relation of time, and (b) that, as 
Kant himself admits, objects of outer sense are, just so 
far as sensible, subject to the form of inner sense. 

' Time,' says Kant, 1 ' is the formal condition a priori of all 
phenomena whatsoever. Space, as the pure form of external 
intuition, is limited as a condition a priori to external 
phenomena alone. On the other hand, because all represen- 
tations, whether they have or have not external things for their 
objects, still in themselves, as determinations of the mind, 
belong to our internal state ; and because this internal state 
is subject to the formal condition of the internal intuition, 
that is, to time ; time is the condition a priori of all phe- 
nomena whatever, the immediate condition of all internal, 
and thereby the mediate condition of all external, phenomena.' 

The difficulty is that e formal condition of all phenomena 
whatever' is incompatible with 'condition of all external 
phenomena.' In order to be in space, in order to limit each 
other, phenomena must be coexistent, not successive. 
' Phenomena ' cannot be used in the same sense when 
qualified by the term ' internal ' as when qualified by the 
term ' external.' If time is the condition of all phenomena, 
what is there in an external phenomenon over and above its 
phenomenality that time should only be its condition 
mediately, and how is this condition any more reconcileable 
with the spatial condition of external phenomena, for 
being merely ' mediate ' ? When Kant speaks of space as the 
form of outer, time of inner, sense, he is understood to mean 
that there is a distinction between outer and inner sense 
independently of that which these forms severally constitute. 
In the sentence ' space, as the pure form, &c.,' if for ' external ' 
we substitute 'in space,' we get a mere tautology. He 
appeals, in effect, to a supposed knowledge of what ' external 
phenomena ' means, in order to explain the statement that 
space, not time, is their condition qua c external.' The ques- 
tion then arises, to phenomena, as such, has the distinction of 
inner and outer any application ? What phenomena are outer, 
what inner '? Clearly to sensations as such, i.e. to sensations as 
they are for a merely sensitive consciousness, the distinction 
cannot apply. They are neither inner nor outer, because 
such consciousness, having no conception of subject or object, 

> Krit. d. r. V., p. 67 ; Tr. p. 30. 


could not refer sensations to either. Can the distinction, 
then, be made to correspond to that between sensation and 
perception? Is feeling inner sense, perception outer? 
We have seen that sensation as such is neither inner nor 
outer, nor could it be held that the sensation of colour was 
inner, the perception of a coloured object outer ; but 
if by 6 feeling' is meant a datum of consciousness which a 
thinking subject regards as a modification of itself in 
distinction from one which it regards as representing the 
quality of an object, and which is called a ' perception,' then 
this distinction may properly be called that between outer 
and inner sense, though the source of the distinction 
would have still to be explained. It clearly is not a dis- 
tinction which originates in phenomena as phenomena, or 
in their relation to sense. It arises out of an intellectual 
interpretation of phenomena, an interpretation by the 
thinking principle that yields alike the 'transcendental object' 
and the 'transcendental subject.' It remains to be seen 
whether, understanding outer and inner sense in this way, 
i.e. as a distinction between the consciousness of objects as 
given (whether perceived or remembered, whether given 
through the senses of sight and touch on the one side, or 
those of hearing and smelling on the other) and the con- 
sciousness of changes in my state, there is any propriety in 
calling space the form of outer, time that of inner sense. 

102. Time, then, is the form of a relation between all 
feelings as attended to, as converted by the presence of 
thought into felt objects determined by mutual relation. 
Mere feelings, i.e. feelings as they are for a merely feeling 
subject, are not so determined, and thus are not in time at 
all. Thus all felt objects, as such (as feelings objectified or 
attended to), are in time or successive, and thus are not in 
space or coexistent. They are only in space so far as by a 
further act of thought they are taken out of this relation of 
succession, which is the primary condition of sensibility, and 
held together as relatively permanent, as one lasting as long 
as the other. This further act has to be performed in order 
to the seeing and touching of objects, and goes to constitute 
perception ; it has not to be performed in order to the hearing 
of them. What happens in ordinary cases of seeing an object 
is, that certain sensations recall a conception already formed, 
and it is impossible to say how little combination of sensations 

TIME. 257 

there need be in order to such recollection. But in the case 
of seeing an object for the first time, unlike anything already 
conceived, I should have successively to attend to sundry 
sensations, and over and above this to hold together these 
successively felt objects in juxtaposition as parts of a space. 
In listening to music, though there is a synthesis of sensations 
by thought, it is not a synthesis of this sort, but one in which 
the sensations, though determined by relation to each other, 
remain successive. 

By detachment of the relation of time or succession, or 
by considering objects as merely qualified by it in exclusion 
of all other relations, we get mere time as an object, which, 
for the reasons above given, may be treated as a quantum. 
The idea of number is not really derived from that of time ; 
it must have been got by abstraction (as explained above) 
of the relation of otherness, before time could be regarded as 
a quantity, before moments could be counted, as they must 
be in order to afford the appearance of being the source 
from which the abstraction of number is derived. 



(Mill, Book II., Chapters V., VI., VII.) 

103. 'The points, lines, circles, and squares, which any 
one has in his mind, are (I apprehend) simply copies of the 
points, lines, circles, and squares which he has known in his 
experience.' 1 So far a clear-headed Kantist would quite 
agree with Mill, except that for e copies of ' he would rather 
say ' identical with,' and for ' has known ' ' knows.' He 
could not hold, as Mill supposes certain nameless persons 
to do, that geometry has ' nothing to do with outward 
experience,' for his doctrine is that geometry has to do just 
with that condition of experience which renders it an out- 
ward experience. The real question is, not whether or no 
geometry c has to do with outward experience,' but (a) what 
is that in experience from which the primary truths of geo- 
metry are derived, and (b) how they are derived from it. If, 
as appears from certain passages, Mill's answer to (a) is 
' sensation,' and to (6) e generalisation ' (as a process by 
which, having frequently observed a ' relation between pheno- 
mena,' we come to assert it universally), then on each of 
these points he is at direct issue with the Kantist, whose 
answer to (a) is, not ' sensation,' but a relation by which 
sensations are determined in becoming perception, but which 
is constituted by the intellect ; and to (b), not ' generalisa- 
tion ' in the sense described (a process which can only result 
in a habit of expecting one sensible event on the appearance 
of another with which it has been constantly associated, a 
habit with which mathematical certainty has nothing in 
common), but a detachment and reduction to its simplest 
expression of the above-mentioned relation, a detachment 
which yields propositions as certain upon the first apprehen- 
sion of the objects to which they relate, as they can become 
upon repeated apprehension. 

For Mill's answer to (a) and (b) see the passages in 

1 Mill, Book IT. chap. v. sec. 1. 


Book II, chapter v, §§ 1, 4, 5, and 6. 1 These can only be 
taken to mean that a straight line is a sensation or collection 
of sensations, that we have repeatedly seen or felt two such 
lines not enclosing a space, and that in consequence by 
generalisation we conclude that they never do or can enclose 
a space. In like manner, that a curved line is a sensation 
or collection of sensations, and that having repeatedly seen 
such a line, when so drawn as to be at every point equidis- 
tant from a certain point, returning into itself or enclosing a 
space, we infer that it always is so, and frame the proposi- 
tion which stands as the definition of a circle. To the same 
purpose is the statement that a point is the minimum visibile, 
the smallest spot of colour that can be seen : a line, we 
must suppose, is the smallest strip of colour that can be 
seen. 2 

104. But then it appears (ib. § 1) that there neither exist, 
nor can be seen or felt or conceived, objects having such pro- 
perties as the definitions state. How, then, can these be 
derived by generalisation from what exists, and is seen and 
felt ? A generalisation is understood to be an inference to 
the uniform occurrence of what has occurred often and 
without exception (that Mill so understands it in this con- 
nection appears from his language (ib. § 4), 'experimental 
proof crowds in upon us in such endless profusion, and 
without one instance in which there can be even a suspicion 
of an exception to the rule, &c.') ; but these generalisations 
would seem to be inferences to the uniform occurrence of 
what has never occurred. 

1 Especially, § 5, 'the exact resem- is the understood limit between lines, 
blance of our ideas of form to the the line the understood limit between 
sensations which suggest them : ' § 6, surfaces. Only through interpretation, 
1 1 have already called attention to the through the new character which it 
peculiar property of our impressions of takes by means of this relation, does a 
form, that the ideas or mental images spot of colour become a point, which, 
exactly resemble their prototypes, and not consisting in the spot of colour, 
adequately represent them for the does not cease to exist when the spot of 
purposes of scientific observation : ' § 4, colour disappears. To call the point the 
'the proposition, Two straight lines minimum visibile is misleading in two 
cannot inclose a space ... is an in- ways: (!) it suggests the notion that 
duction from the evidence of our senses:' sensation constitutes the point, (2) it 
§ 1, 'nothing remains but to consider implies that the point, as a quantum 
geometry as conversant with such lines, {minimum is a term of quaDtity), is not 
angles, and figures, as really exist; and divisible; that the possible division of 
the definitions must be regarded as some magnitudes ceases where visibility (with 
of our first, and most obvious general- whatever eyes or instruments) ceases, 
isations concerning those natural ob- In truth every quantum is divisible, 
jects.' since it is not a quantum unless made 

2 The true account is that the point up of parts, 

s 2 


We have never seen two straight lines, having met, 
continuing to diverge, or a line, so drawn as to be at every 
point equidistant from a certain point, returning into itself, 
for we have never seen such lines at all. What basis have 
we, then, for the generalisation? (a) How do we get 
materials for making it, and how can it be made ? (b) How 
can it apply to reality when it is made ? Generalisation is 
a process by which from a multitude of similar singular 
judgments we pass to one universal judgment; from 'this 
a7id this and this man dies ' to ' all men are mortal.' Now 
this is clearly not the process by which we arrive at any of 
the general propositions in Euclid that are said to be ' proven.' 
We have never observed the square on the hypotenuse to be 
equal to the squares on the containing sides before we take it 
as universally true. To answer [a) from Mill's point of view, 
it must be supposed that, seeing innumerable lines nearly 
straight, we in each case suppose them to be quite straight, 
and then generalise from these suppositions. In that case 
the generalisation, which according to Mill is ' faultless as a 
generalisation,' i.e. is perfectly warranted by the instances 
from which the inference is made, is not really from the 
'evidence of the senses,' but from something which we habi- 
tually substitute for this : geometrical forms are not, as Mill 
says, ' copies of impressions,' but alterations of them. How 
is such substitution and alteration to be accounted for? 
How can I see a line to be nearly straight, if I have no prior 
idea of straightness ? How is it that a mind, merely passive 
and receptive, having no antecedent conception of a straight 
line, nay (for so Mill says), unable to conceive it, yet habi- 
tually assumes such a line in place of that which it sees, as 
it must do if the definitions are to be explained as the result 
of generalisation? The difficulty is only evaded by saying 
that we get the objects from which we generalise the defini- 
tions by abstraction from what we see, i.e. by attending 
exclusively to some part of what we see. We cannot 
abstract what is not there to be abstracted. If straight lines 
are not in the things which we see, by no abstraction or 
exclusive attention can they be got out of them. In short, 
if we admit that the primary mathematical truths are got 
by generalisation, we are thrown back on a prior question as 
to abstraction. Is it a process by which we arrive at the 
primary propositions about straight lines and circles? If it 


is, it must be supposed to be exclusive attention to certain 
qualities of a * complex phenomenon.' But straightness, 
according to Mill, is not a quality of any line. Having the 
idea of a straight line we may no doubt, on seeing a line not 
quite straight (which implies this idea), make abstraction of 
its imperfection and regard it as straight ; but this does not 
explain how we came by the idea of a straight line. In 
truth there is no more difficulty about seeing a straight line 
than about seeing a line, or rather there is less. 1 A straight 
line is constituted by the simplest form of that intellectual 
act which is necessary to ' seeing ' any line whatever. 

105. As to (b), Mill says that in applying to the real 
world conclusions derived from any set of geometrical pro- 
positions, we correct them by a ' fresh set ' of propositions. 
But the * fresh set ' is either one derived from the same unreal 
assumption (in which case it is not easy to see how one fiction 
should be corrected into truth by another fiction), or else 
6 relates to physical and chemical properties of the material,' 
in which case the correction does not imply that such and 
such a geometrical figure does not ' really exist ' or has not 
the properties which geometry asserts of it, but merely that 
there are physical and chemical conditions which prevent 
any given material from retaining such a figure. In other 
words, the hypothesis which it corrects is not that there 
are geometrical figures which really there are not, but that 
things are subject to geometrical conditions only (which, as 
we have seen, is the true account of the geometrical hypo- 
thesis). And this is the way in which Mill himself states 
the matter at the end of § 2 of Chapter V, 2 without seeing 
tha/t it is quite a different account from that which he has 
previously given, according to which straightness is not 
a property of any lines whatever, or circularness (as defined) 
of any curves whatever. 3 

1 Do we see straight lines and circles, 3 According to Mill, mathematical 

according to Euclid's account of them ? truths are ' approximately true ' of 

Do such straight lines and circles really nature or reality. What does this 

exist ? In whatever sense we see lines, mean ? When we say that a generali- 

wesee straight lines; in whatever sense sation is approximately true, we mean 

lines really exist, straight lines do so. that it is true in most cases, but not 

1 ' If the hypothesis merely divests in all. But a mathematical definition, 

a real object of some portion < f its says Mill, ' is not exactly true in any 

properties, without clothing it in false case.' But it is 'nearly true' in all. 

ones, the conclusions will always ex- Yes, but the transition from the Q< arly 

press, under known liability to correc- true to the exactly true is just the 

tion, actual truth.' difficulty That so and so is nearly 


106. Mill, taking advantage of Wliewell's misleading 
statement that ' intuition ' = ' imaginary looking,' opposes 
to the view that mathematical truths are known by such in- 
tuition his own view that they are known by 'experience ' 
or ' real looking.' 1 But how can this be when, according to 
him, we never see, and there really are not, figures correspond- 
ing to the definitions ? 2 As against Whewell he is quite 
right on this point. 3 If they are not known by real or 
sensuous looking, they will scarcely be known by ' imaginary,' 
which is only reproduction of the ' real ' looking. In what- 
ever sense they are known by 'imaginary,' they are known 
also by ' real,' looking. In truth they are known by an act 
of intellectual synthesis, involved alike in ' real ' and ' imagi- 
nary' looking, and (in their 'purity') by detachment from 
sensible conditions of the element which this act constitutes 
in all looking. The question whether we can imagine a sur- 
face without colour, or length without breadth, &c. is very 
unprofitable. We can think of surface without colour, and 
mere surface represents a certain condition of reality in 

' Dr. "Whewell ' (says Mill) ' thinks it unreasonable to con- 
tend that we know by experience that our idea of a line 
exactly resembles a real line. * It does not appear,' he says, 
' how we can compare our ideas with the realities, since we 
know the realities only by our ideas.' We know the reali- 

true, must mean either that it is so in 8 With Kant ' intuition ' is not ' ima- 

the majority of cases (it is not so ginary looking.' It includes imagina- 

with the circle), or that it -would be tion, but also perception; and space 

true but for qualifying circumstances. being the form of the one as well as of 

Mill would say that the latter is the the other, the properties which geometry 

case with the definition of a circle. If predicates of space apply equally to 

so, there is such a thing as a circle perceived and to imagined objects, 

according to the definition, and we Kant, however, does identify ■ pure ' 

have a notion of it, only it is always (or non-sensuous) intuition, which is 

de facto qualified by something else. employed in geometry, with • pure ima- 

Mathematical figures, then, are in what gination,' but then this is imagination 

we experience. If so, there is an ele- without the essential characteristic of 

ment in our experience which is not imagination, viz. reproductiveness. For 

feeling, immediate or reproduced, and Kant's ' pure intuition ' as exercised in 

which can be separated from all other geometry, it would be better to substi- 

elements, though never constituting a tute ' conception of the pure element in 

separate impression. intuition.' It is important to notice, 

1 Book II. chap. v. sec. 5. however, that with Kant ' intuition ' 

2 According to Mill (ib. sec. 5), carries no antithesis to what Mill 
axioms about straight lines (as well as reckons experience of real objects, 
axioms of quantity) are ' true without What Mill says that we ' see,' Kant 
any mixture of hypothesis.' How can says that we ' intuite,' and that this 
this be when straight lines only exist intuition, though involving sensation, 
by hypothesis ? involves a ■ pure ' element as well. 


ties (I conceive) by our eyes. Dr. Whewell surely does not 
hold the "doctrine of perception by means of ideas," which 
Reid gave himself so much trouble to refute.' l Whewell 
would have been quite right if he had said that the supposed 
comparison is absurd, since the idea is the reality, — since it 
would be a comparison of what remains in the mind from 
reality with what the mind has put into reality, or has added 
to sensation in order that ifc may become reality. To com- 
pare the idea of a line as expressed by the definition with a 
sensation of colour, is nonsense; to compare it with the 
boundary of a coloured surface is to compare it with itself. 
It would be easy to retort on Mill that, according to him, 
we do not know the realities in question (mathematical 
figures) at all. He would be quite right, however, in saying 
that we know them ' by our eyes ' just as much as by ' imagi- 
nation.' The ' doctrine of perception by means of ideas,' 
which he refers to as refuted by Reid, is the doctrine of ' re- 
presentative ideas ' ; the doctrine that we cannot perceive 
material things in themselves, since ' mind ' and ' matter ' 
are wholly alien to each other, but that we perceive ' ideas ' 
of them, a tertium quid which is neither matter nor mind. 
Something very like this was held by Locke. ' Idea ' (ac- 
cording to him) is the immediate object of the mind in per- 
ception and thinking. Some ' ideas ' are copies of qualities 
of matter, others are effects, though not copies, of such quali- 
ties. Mill himself is by no means free from it. It is implied 
in all language which speaks of us as knowing things through 
sensation, and describes sensations as impressions made by 
external matter. It never can be entirely got rid of save by 
the recognition of ' matter ' as a congeries of relations con 
stituted by thought. When Whewell talks of a ' reality 
known only by our ideas,' he is obscurely thinking of matter 
in this way. The worst of him is that he is still hampered 
by the notion that there is one sort of reality known by our 
* ideas,' another by our ' eyes.' 

107. The assertion of the reality of space, of space quali- 
fied according to geometrical definitions, is not to be under- 
stood as if space were absolute, fixed and final, a property of 
things standing over and against the spirit. As we have 
seen, it is a relation which thought constitutes, and one which 
is the negation of another equally necessary relation, that of 

1 II. v. see. 5, note. 


time, and which itself carries with it the negation of itself, 
since objects outside each other (or in space) are yet not out- 
side each other, since each (by its outsideness) makes the 
other what it is. ' Body cannot act where it is not,' yet it 
determines that which is where it is not by simply not being 
there. Mill thinks that whereas physicists were long misled 
by the supposed inconceivability of a body acting where it 
is not, they have now found out that it does act where it is 
not, which shows that the ' inconceivable ' may become 
conceivable, and merely means that which has not yet been 
discovered. But in truth the doctrine in question, though 
perfectly true in itself, carried with it its own negation. The 
whole truth consists of it and its negation. It represents the 
abstract view of space, i.e. the consideration of one object 
merely as outside the other, without the complementary view, 
that of the external objects each by its outsideness makes the 
other what it is, and is thus not outside it. Body ' cannot 
act where it is not,' but it is where it is not. 

108. All general truths are necessary truths ; only the 
whole of truth is truth of which the negation is inconceivable. 
It is misleading to oppose ' necessary truths ' and ' truths 
of experience.' There are no truths which * rest on evidence 
of a higher and more cogent description than any which 
experience can afford,' l according to the right view of 
' experience.' When Kant said that mathematical truths, 
being necessary, could not be derived from experience, he 
meant (in effect) experience as Locke understood it. He 
meant that they were not the result of generalisation as the 
process by which, from observation that two phenomena 
often and without exception accompany or follow each other, 
we arrive at the judgment that they always do and will. 
This is what Mill himself generally understands by the 
derivation of mathematical truths from experience. It is a 
process which, as Locke and Hume were quite aware, can 
only yield a habit of expectation, of various degrees of 
strength, and can as little explain the certainty of a judg- 
ment founded on a crucial experiment in physics as it can 
that of a mathematical proposition. No general truths 
about nature are really got in this way. They are not 
summaries of events which have happened very often, and 
are so far likely to happen again. According to Mill's own 

1 lb. II. v. bee. 6. 


account, they are got by the analysis, according to the ' in- 
ductive methods,' of an experience regulated by the concep- 
tion of the uniformity of nature. If they were summaries of 
events which have happened very often, and are so far likely 
to happen again, they might cease to be true any day. A 
feeling might occur in a sequence in which it has not pre- 
viously occurred, and thus the proposition previously taken 
to be true might have to be reversed. But such summaries 
are not general truths at all. Just so far as propositions 
about nature are general and true at all, they are necessarily 
true. They represent the relation of a phenomenon to its 
conditions, and this relation, on the principle that the world 
is one (a principle without which there is no knowledge at 
all), cannot vary. If the relation of the phenomenon to its 
conditions is misapprehended, if the supposed conditions of 
the phenomenon are not really conditions of it, then the 
propositions in question are not true. The case is different 
when the supposed conditions of a phenomenon are really 
conditions of it, but are subject to modifying conditions. In 
this case it may properly be said that the corresponding 
judgment is ' nearly true.' Most scientific judgments, I 
suppose, are of this sort ; and just so far as they are true, 
they are necessary. The right account of the matter is not, 
that the phenomenon is but might not be dependent on the 
known conditions, but that it is dependent on known con- 
ditions in conjunction with others still unknown, an alteration 
in which, apart from any alteration in the known, might 
produce a different phenomenon. The qualification of the 
necessity of the judgment is also a qualification of its truth. 

The distinction of mathematical propositions from these 
propositions about nature, which are ' nearly true,' arises 
from the fact that the properties of space are not dependent 
on other conditions, as the conditions to which a phenomenon 
is referred in a scientific judgment which is ' nearly true ' 
depend on other conditions. The possession (say) of circular 
figure by any particular body at any particular time depends 
of course on most complex conditions, but a circle may be 
a circle without having any properties but those which 
depend on its circularity. Hence there is no place for 
' approximate truths ' in geometry. 

109. To say that a general proposition is true, and to 
say that its contradictory is inconceivable, are one and the 


same thing. The moment I conceive the general proposition 
as true, I cease to be able to conceive its contradictory as true. 
Whewell is quite right in saying that, the law of chemical 
combination in definite proportions once being conceived to be 
true, its contradiction becomes inconceivable. The contradic- 
tory would be conceivable only if the law were not conceived 
as wholly true. Given any proposition conceived as wholly (un- 
conditionally) true, you cannot conceive its contradictory to 
be true consistently with that idea of the unity of the world 
without which no proposition could be conceived to be really 
either true or untrue. The proposition in question is, I sup- 
pose, conceived as wholly true, because the chemical elements 
are considered ultimate, and thus not subject to any further 
conditions which could account for a variation in the propor- 
tions in which they combine. If the combination of these 
elements should be found to depend on any ulterior cause, 
then a variation in the proportions in which they combine, 
if it could be accounted for as the result of this cause, would 
be compatible with that unity of the world which is the 
condition of knowledge, i.e. it would no longer be a 
matter of chance. It is not variation, but unaccountable 
variation, that is inconceivable. The fact that the law was 
only lately discovered is nothing against the inconceivability 
of its contradictory when discovered. Unless all unconditional 
truths are truths that all men have known from the beginning, 
it is absurd to answer the doctrine that the contradictory 
of a certain truth is inconceivable by saying that the truth 
itself was not always conceived. The question is whether 
any one, having conceived the truth and the grounds on which 
it rests, can at the same time conceive its. contradictory to 
be true. On the other hand, the inconceivability of its 
contradictory is no independent test that a proposition con- 
ceived as unconditionally true is really true, since it is 
the same thing with the conception of it as thus true. The 
proposition is not true because its contradictory is inconceiv- 
able. A general proposition is true because it is the only 
way of explaining the facts to which it relates, compatibly 
with the unity of the world ; which implies that its contra- 
dictory is inconceivable. The inconceivability of the con- 
tradictory may be only provisional, but that is because the 
conception of the proposition as true is only provisional. 
A new discovery might render the contradictory of the given 


proposition conceivable, but only by rendering the proposition 
itself untrue. Where the truth is ultimate (as the truth of 
the above-mentioned chemical law must be ultimate so long 
as chemistry remains a separate science, since it affords the 
only way of holding the chemical facts together in a science, 
unless they be accounted for by some other science), the 
inconceivability of the contradictory is absolute. 

110. Thus Mill is incidentally right in denying that the 
inconceivability of the contradictory is the test of truth, but 
shows that he is not right on right grounds by holding that 
a proposition may be conceived true without its contradictory 
becoming inconceivable. By ' conception ' Mill generally 
understands a ' mental picture,' by * inconceivable ' that of 
which a mental picture cannot be formed. Now the pro- 
gress of science clearly does not affect our power of forming 
mental pictures, or, if it affects it, increases, not diminishes, 
it. Thus the progress of science cannot render anything 
inconceivable, in this sense, that was not so to begin with. 
That alone is thus 'inconceivable' (or, to adopt Mill's 
phraseology, 6 unimaginable ' as distinct from ' unbeliev- 
able ') which cannot be presented as in space. A last point 
of space, to take Mill's instance, is unimaginable because 
incompatible with the conditions of intuition ; it is a space 
(i.e. something determined by outsideness), which yet has 
nothing outside it. And in this instance what is unimagin- 
able is also inconceivable, because a space (a point or piece 
of space) is nothing if not an object of intuition, and here 
the conditions of intuition are denied. Mill apparently 
holds l (a) that nothing can be true which is unimaginable 
in this sense, but (b) that plenty of propositions may be true 
without their contradictory being unimaginable. He is 
right in (6), wrong in (a). Wrong in (a) because relations, 
which are our principal concern in knowledge, are not 
imaginable. None of the propositions which state laws of 
nature state what is imaginable, though the laws which 
they state relate to imaginable objects. Motion in all its 
forms is unimaginable, 2 and all laws of nature are in a wide 

1 Book II. chap. vii. sec. 3. points and lines, but not the transition 

2 Take the proposition, ' The radii from one to the other. Mill would say 
vectores of planets and comets traverse that we can imagine what we see, and 
equal areas in equal times.' Admitting we see bodies move. But we do not 
that ' planets ' ' radii ' and ' areas ' are ' see ' them move (even in the sense of 
imaginable, at any rate the traversing ' intuition ') : motion is that by which 
of the area is not. You may imagine we explain what we see. 


sense laws of motion. Therefore Mill is only right in (b) 
because he is wrong in (a). The reason why propositions 
may be true without their contradictories being unimaginable, 
is that the propositions themselves relate to the unimagin- 
able, or, more properly, to that in regard to which the 
question whether it is imaginable or no is unmeaning. 

111. Mill admits another improper sense of ' inconceivable ' 
as = unbelievable. 1 I do not find that he anywhere exactly 
explains this, but he implies that belief is the result of past 
sensations, and that its strength is proportionate to the 
number of these, and to the uniformity in the order of 
their occurrence. Sensations ' register ' themselves (to use 
Herbert Spencer's language), and belief represents accumu- 
lated entries in the register. In short, he has no other 
notion of belief than that of Hume, that it is the involuntary 
return of an idea with such liveliness as almost to = an 
impression. If * inconceivable ' = ' unbelievable ' in this 
sense of belief, it is quite true that 'inconceivableness is an 
accidental thing, not inherent in the phenomenon itself, but 
dependent on the mental history of the person who tries to 
conceive it.' 2 In this sense of ' inconceivable,' the difficulty 
is not to show that the contradictory of scientific propositions 
is conceivable, but to show that these propositions them- 
selves are conceivable, requiring us, as they commonly do, to 
set aside a ' belief,' founded on a long succession of sensible 
events, for a theory of which either the senses give no evi- 
dence at all or which is founded on the result of a single 
experiment. So far as we have the ' evidence of our senses ' 
for anything (which we have not), we have it for the motion 
of the sun and against the motion of the earth. Why do 
we 'believe' the contrary? Mill admits 3 the difficulty of 
conceiving (in his sense) the contrary, but recognises none 
about the belief; but if belief is an 'accidental thing,' 
dependent on the past sensitive experience of the indi- 
vidual, the Copernican theory is quite as ' unbelievable ' as 
1 inconceivable.' 

112. Upon Mill's doctrine, that things really exist quite 
independently of thought or conception, and that the latter 
merely results from impressions which things make on us 
through sensation, neither the Copernican nor any other 

1 Book II. chap. vii. sec. 3; unci chap. 2 Ibid. chap. v. sec. 6. 

v. sec. 6. 3 Ibid. chap. vii. sec. 3, end. 


theory can be accounted for, for every theory corrects sense, 
or rather (since mere sense gives nothing to correct), the 
first inferences from sense. In truth the reality of things is 
their determination by each other as constituents of one 
order, a determination which only exists for thought. It is 
not that there is first the reality of things, and then a theory 
about it. The reality is a theory. No motion is properly a 
phenomenon, but a relation between phenomena constituted 
by a conceiving mind ; a way of holding together pheno- 
mena in thought. Just as the motion of a planet is a way 
of holding together certain phenomena, the only possible 
mode of holding those phenomena together as one, so the 
Copernican system is the only way of holding the planetary 
motions together as one, as changing appearances of one 
principle. It is a reality (not a mere theoiy about reality), 
but it is a conception, though a conception which any one of 
us may or may not have made his own, may or may not 
conceive or believe : and the ' inconceivableness of its con- 
tradictory ' is not an ' accidental thing, dependent on the 
mental history of the person who tries to conceive it,' but 
' inherent in the phenomena ' which form the system, — not 
indeed as separate phenomena (for as such they have no 
reality), but as a system. 

113. As against Spencer, Mill is quite right, for Spencer 
has no other notion of the ground of belief than Mill's ; and 
to make ' inconceivability of the contradictory,' thus under- 
stood, the measure of what can be true, is to measure truth 
by involuntary habits of expectation and memory, as they 
happen to stand in any individual or set of individuals at 
any time. In fact, Spencer's ' inconceivables,' though in the 
passage quoted by Mill he tries to represent them as ' un- 
believables' in the sense of that which accumulation or 
uniformity of experience prevents us from believing, are 
negations of certain formal ideas, which in truth carry with 
them their own negation. The negative, supposed incon- 
ceivable, so far from being so is the necessary complement 
of the conception to which it is opposed. We ' believe in 
our own sensations ' because c the negation of this belief is 
inconceivable.' We believe that ' space, time, force, exten- 
sion, figure are objective realities,' because 'we cannot by 
any effort conceive these objects of thought as mere states 
of our mind ; as not having an existence external to us.' 


What is meant by 'belief in our own sensations ' ? It must 
be something different from the sensations themselves. 
Spencer seems to mean by it the reference of sensations to 
an external cause, the thought of them as representing ' ob- 
jective reality.' There is nothing, however, in their constant 
recurrence, which Spencer seems ordinarily to regard as the 
source of belief, to produce this reference. No one has ever 
been able to show why a sensation, which does not at first 
refer itself to an external cause, should come to do so on 
repetition. The Humian explanation would be that this 
reference to an external cause means the involuntary expec- 
tation that other sensations will follow, an expectation which 
the constant conjunction of these other sensations with the 
given one will account for. Such expectation may, no doubt, 
be thus accounted for, but the question is whether such ex- 
pectation can account for science ; whether the conception of 
objective reality, as thus reduced to involuntary expectation, 
can afford the basis which the possibility of science pre- 
supposes. Spencer strongly rejects the Humian view, but 
has really nothing to put in its place but a long-winded 
version of Dr. Johnson's refutation of Berkeley, which con- 
sisted in kicking a stone. Not seeing that the belief in 
question is the reference of sensations by thought to an 
object which itself constitutes, he regards the externality of 
the ' sensible thing ' (the cause to which sensations are 
referred) as outwardness to the mind of a kind with the out- 
wardness of one space to another. He makes what is in 
fact but our first thought about the world a final and absolute 
truth, of which the negation is inconceivable. Sensations 
are 'objectively real,' no doubt; yet nothing is what it 
seems. A sensation has its objective reality just in that 
which, as a separate sensation, it itself is not, viz. its relation 
to the universe of things. It is produced by an ' external ' 
cause no doubt, but this externality of the cause, if it 
means (a) externality to the mind, is externality which the 
mind itself constitutes, and is thus within the mind ; if it 
means (b) externality of one thing to other things, is an exter- 
nality according to which each of the externals is the essence 
or the qualifying nature of that to which it is external, and 
is thus within it. 

114. The appearance of our inability to ' conceive ' space, 
&c. 'as not having an existence external to us,' arises (a) 


from a confusion of the objective with the external, (b) from 
our misinterpreting relations which the thinking subject 
constitutes as if they were relations under which thought 
itself exists, (a) J Space, &c.' are ' objective ' — not modes of 
our consciousness — in the sense that they do not depend on 
the consciousness of any one in particular (as, e.g. the com- 
position of Beethoven's symphonies depended on the con- 
sciousness of that particular individual, as the Christian 
religion depends on a consciousness shared by a multitude of 
persons, and as the perception of colour depends on the con- 
sciousness of all possessed of certain sensitive faculties), but 
on the universal (though by no means exclusive) conditions 
under which consciousness regards its object. But in this 
there is no externality. The object of consciousness is not 
external to the subject, or, if it is, it is external in the way 
that carries its own negative (as explained above), (b) Ex- 
tension and force are relations implying the externality to 
each other of the things related. Externality is predicable 
of these things. But the relations themselves are not ex 
ternal to anything. Externality is not predicable of them ; 
still more clearly it is not predicable of the thinking subject 
which constitutes them. Thus they, only seem external to 
thought through the fallacy of treating the mind itself as 
one of the things between which the relation of externality 

115. 'Two and one are equal to three,' according to 
Mill, 1 ' is a truth known to us by early and constant ex- 
perience ; an inductive truth ' ; i.e. it has happened to us 
very often and without exception to find that the same 
three objects, which as presented in separate parcels have 
produced a certain set of sensations, will produce another 
set of sensations if presented in one parcel, and vice versa. 
We are thus led to believe that this always will happen in 
the case of any three objects whatever, a belief which we 
state in the form c 2 + 1 = 3.' 

This explanation assumes that which is to be explained, 
viz. the act of counting. It assumes an aggregate counted 
as three : otherwise we could never have compared sensations 
produced by one parcel of three objects with those produced 
by the same objects in two parcels. Having counted the 
objects as three, we already know that 2+1 = 3, since we 

1 Book If. chap. vi. sec. 2. 


have only got them by adding one to one so as to form a 
number, and then adding one to this number two. The 
whole question at issue relates to the synthetic act of 
counting, the act by which various numbers are given. 
Given any number (e.g. twelve), all propositions which state 
equality between it and various combinations of its elements 
are derived by mere analysis (e.g. 8 + 4=12; 7 + 5 = 12). 
How, without sheer nonsense, can the mental act which 
yields the number twelve be called a * generalisation from 
experience,' i.e. an assurance gradually obtained that what 
has happened constantly will always happen ? ' Are there 
not/ it may be said, ' collections of objects amounting to 
twelve which are always striking our senses?' Exactly so, 
but do these amount to twelve for a merely sensitive con- 
sciousness ? e We see twelve pebbles ; twelve pebbles strike 
our sense of sight in a particular manner.' Is then twelve 
a light or a colour 9 In truth a visible object ' impresses 
the senses thus ° °,' l only because in seeing it we count ; 
only because we either hold together successive sensations 
in one compound of parts, or think the sensation of a 
moment into parts which we proceed to add. In short, only 
because c sensible objects ' are already numbered or numer- 
able through a synthetic act of thought, is there an appear- 
ance of our deriving ideas of number from them by abstraction 
and generalisation. 

1 Book I. chap. vi. sec. 2. 



(Mill, Book II., Chapters I., II., III.) 

116. In the theory of syllogism, as in that of ' necessary 
truths,' Mill is to a great extent right as against the doctrine 
which he attacks. If the question at issue is whether the 
6 ground,' on which we ' draw conclusions concerning cases 
specifically unknown to us,' is best stated by saying that 
' the unknown case is proved by known cases,' or by saying 
that ' it is proved by a general proposition including both 
sets of cases, the unknown and the known,' l Mill's answer, 
as here given, is quite right. It is no doubt absurd to say 
that the mortality of Socrates is proved by the truth that 
' all men are mortal.' But the question remains whether 
the proof that Socrates is mortal is that this, that, and the 
other man have died (as Mill holds, chap. iii. § 4) ; whether it 
is proved by ' generalisation from observed particulars given 
by sense,' of which the general proposition is a summary 
statement, in the sense in which Mill understands it. In 
short, the question at issue concerns the nature of ' gene- 
ralisation.' Is it (a) a process from concrete individuals, by 
omission of their distinguishing attributes, to a class ; or (b) 
a process from a constantly observed sequence of one sen- 
sible event on another to the involuntary expectation of one 
upon the recurrence of the other; or (c) a process from 
a multitude of separate events to their uniform conditions 
(relations) or single cause? 

(a) is the doctrine of the scholastic logic, of which the 
history was determined by Plato's original failure to dis- 
tinguish between alaO^rov as mere sensation (sensation as it 
would be for a merely feeling consciousness), and alcrdrjrov as 
a concrete sensible thing (a complex of attributes constituted 
by relations). alaO^rov in the former sense (or alaOr^rov 
in the sense of our first imperfect interpretation of sense), 

1 Note to Chapter VI. 
VOL. II. . t 

274 LOGIC : MILL. 

but not in the latter, may properly be opposed to vor^rov or 
ettlctttjtov. From the opposition of alaOijrov in the latter 
sense to votjtov arises the doctrine that on the one hand 
things in their definite concrete individuality are given to us 
in sense independently of any action of thought, and on the 
other hand that the action of thought or the process towards 
knowledge consists in a gradual abstraction from, or omis- 
sion of, the distinguishing properties of the individuals, till 
a universal, in the nature of a class possessing some fraction 
of the attributes of the individuals, is arrived at. Of such 
a class one of the attributes included in the connotation of 
the corresponding name was predicated in the proposition 
forming the major premiss of a syllogism. On its being so 
included, and on the inclusion of the individual or species 
forming the minor term within the class (i.e. on the pre- 
dicability of the class-name of this individual or species), 
the validity of the syllogism depended. Locke was right, 
once for all, in saying that syllogism could only be analytical 
of a 6 nominal essence.' 

117. Mill retains syllogism as representing a real inference, 
but holds that the conclusion is drawn, not from the general 
proposition, in the assertion of which, as he admits, the 
conclusion is already asserted, but from the ' particulars ' of 
which this general proposition is a register. But his view 
is open to virtually the same objection as the old view of 
syllogism. Is the c particular ' of which an attribute is 
asserted in the conclusion one of the particulars which have 
been already observed to have this attribute (the particulars 
of which the middle term is the summary), or is it not ? If 
it is, then there is no inference to it. The conclusion is just 
is clearly involved in the data, these being the observed 
particulars, as it is in the general proposition according to the 
old view. If it is not, how is the inference justified ? How 
is the inference valid unless the sir ay coy rj is Sea irdvrcov ? and 
if it is 8ta Trdvrcov, how is it inference at all ? 

One answer to this objection is to adopt Locke's denial 
of ' general certainty ' in regard to nature ; to say that what 
we call so is an involuntary habit of expectation which on 
the recurrence of a sensation recalls the idea of its usual 
attendant with great vivacity. If generalisation were what 
Mill usually describes it as being, inference that what has 
happened constantly will happen always, this would be the 


only possible account of the certainty which it produces. 
The question is whether this will explain science, which is 
essentially an effort by ' interrogation of nature ' to get 
behind the usual to the uniform. 

118. The other answer is that the inference to the 
mortality of Socrates rests on the observation neither of all 
the particulars nor of many particulars, as all or as many ; 
that it is neither an STraycoyr} Bta ttclvtcov, which is no 
inference at all, nor an STraycoyr) Bid ttoWcov, which can yield 
no scientific certainty; that it has nothing to do with the 
quantity of the particulars, but only with, their kind ; nothing 
to do with how often an event happens, but only with the 
question what it really is that happens in each event. 
Inference is a process from the e ordo ad nos ' or ' ad sensum' 
to the ' ordo ad universum,' from the 6 phenomenon ' in the 
proper sense to its conditions, a process to which the mere 
repetition of occurrences in ordine ad sensum contributes 
nothing. The inference to all possible cases of a like event, 
so far as made at all, is made in the first complete discovery of 
the conditions of the single event. Once know what death 
really is in the case of a single man, i.e. the conditions on 
which it depends, then I learn no more, by seeing any number 
of men die, I do not become any more certain that Socrates 
will die. Whatever uncertainty there may be as to the 
mortality of Socrates consists in the uncertainty whether 
the ascertained conditions of mortality are present in his case 
or no ; whether the resemblance of Socrates to the men who 
have died (or, it may be, to single men who have died, for the 
number of cases makes no difference) is a resemblance in 
respect of the conditions on which mortality depends. No 
doubt, in the process of ascertaining what these conditions 
are, a great number of cases may have to be observed in 
order to the exclusion of unessential circumstances ; but 
the observation of such cases in order to ascertain what 
really happens, what are the conditions of the given 
phenomena in each, is absolutely different from the observa- 
tion which from the constant occurrence of an event leads to 
the expectation of its uniform continuance. The former is 
the sort of observation which Mill has in view when he is 
explaining the inductive methods; the latter is the kind which 
he has in view in his account of the generalisation by which 
from many particulars, registered in the major premiss of a 

T 'J. 


syllogism, we infer the proposition which forms the conclusion 
of such a syllogism. 

119. Mill holds all syllogism to be of that kind which 
the scholastic logic held it to be in cases where the major 
premiss was founded on sTraycoyrj, except that he supposes 
the need for siraycoyrj being Bta nrdvTwv to be in some un- 
explained way dispensed with ; so that the inferred case has 
not been itself one among the cases on which the inference 
is founded. This, however, is to save inference from being 
nugatory by making it invalid, unless the dispensation with 
sTraycoyr] Boa irdvrcov can be justified. c But,' it will be said, 
' Mill's inductive methods are just what enable us to dis- 
pense with i7raj(oyrj Siiz 7rdvro)v. 9 True, but that is because 
they imply a conception of inference in fact wholly different 
from the conception of it as generalisation from many 
observed cases to all possible cases, from what has happened 
often to what will happen always ; a conception according 
to which a general truth is something quite different from 
a summary statement of a multitude of particular events 
(which is what Mill understands it to be in his doctrine of 

' A principle ascertained by experience,' he says, 1 ' is 
more than a mere summing up of what has been specifically 
observed in the individual cases which have been examined ; 
it is a generalisation grounded on those cases, and expressive 
of our belief that what we there found true is true in an 
indefinite number of cases which we have not examined and 
are never likely to examine.' Yes, it is ' more than a mere 
summing up ' because it is not a ' summing up ' at all. Mill 
is governed by the old view that the general proposition 
concerning matter of fact is a summary of what has often 
happened ; at the same time he sees that such a summary is 
not a scientific truth, neither proven nor a basis for proof. 
Accordingly he regards it as a summary of what has been 
observed and something more. In truth, it is only this some- 
thing more because it is not a summary of observed phaa- 
nomena at all, but a statement of the permanent conditions 
of these phenomena, which conditions once ascertained there 
is no further inference from these phsenomena to an < in- 
definite number of cases which we have not examined.' 

' From instances which we have observed, we feel 

1 Book II. chap. i. sec. 3, end. 


warranted in concluding that what we found true in those 
instances holds in all similar ones, past, present, and future, 
however numerous they may be.' l Such a conclusion is 
nugatory. There is no inference from those instances to all 
similar ones. If they are really similar, they are covered by 
the principle discovered in the observed instances. In that 
case the conditions of the phenomenon are the same in all 
cases of its occurrence, and the phenomena are really 
identical, i.e. distinct in time or ad sensum, but one in essence 
(relations) or ad universum. Inference lies, not (as Mill says) 
in the generalisation from observed instances to all, but (a) 
in the discovery of the real conditions of the observed 
instances, (b) in the discovery whether other apparently 
like instances are really like. Given the real similarity of 
the other instances, there is no inference to them. 

120. In geometry there are no c observed phenomena,' no 
ordo ad sensum, at all. Thus instead of (a) above we have 
inference consisting in the combination, primarily, of ele- 
ments consisting in the simplest forms of the limit, and then 
of figures whose properties are known as resulting from such 
combination, and the consideration of what together they 
imply. The conclusion arrived at in this way is in its 
nature universal, though obtained by the construction of a 
single figure, just as the proposition which states the con- 
ditions of a single phenomenon is universal. In each case 
the proposition is not a statement of what happens here and 
now to me ('this feels hot'), but a statement of a relation 
which is not in time at all, a relation which between the 
same things is eternally the same. You analyse a particular 
drop of water into certain proportions of oxygen and hydro- 
gen. You find by means of a particular construction that 
the squares on the containing sides of this right angle to- 
gether = the square on the hypotenuse. In the first case 
you know at once that water is always composed of oxygen 
and hydrogen in the same proportions when the conditions 
are the same as those under which you analysed it (in other 
words, your conclusion is intrinsically universal). In the 
latter case the conclusion is not any more universally 
true, but it is of a different kind, of a kind which renders 
the qualification ' under the same conditions ' superfluous, 
because the conditions cannot be different. For the same 

1 Book. II. chap. iii. sec. 3. 


reason stage (b) of inference, as above stated, has no place 
in geometry. 

Thus Mill is quite right in holding l that inference in 
geometry is not any less 'from particulars to particulars' 
than in natural science, but only because in neither is the 
inference 'from particulars to particulars/ in the sense of 
'from sundry events to another event.' The inference in 
natural science no less than in geometry is to a universal, 
to an eternal and unchangeable relation; and in natural 
science it may be from a single phenomenon, just as in geo- 
metry from a single construction. The difference lies in the 
dependence of the ascertained conditions upon other con- 
ditions in the one case, their independence in the other. 

Just as in geometrical reasoning there is nothing corre- 
sponding to (b) above, so in the lawyer's reasoning there is 
nothing corresponding to (a). The lawyer has his general 
proposition given him by the law. Everything depends on 
making out the particular case to come under the general 
rule, i.e. on a process analogous to (b). 2 Thus geometry is 
constantly arriving at unqualified or unconditional general 
truths, natural science at conditional general truths, the 
lawyer never at any general truths ; he takes all his general 
truths for granted, and shows how particular acts can be 
construed as covered by them. 

121. What Mill gives 3 as the 'universal type of the 
reasoning process,' corresponding to syllogism, does not as 
it stands properly represent either (a) the process, or (6) (as 
syllogism may do) the result of reasoning. As regards (a) it 
is clear that the process of inference depends on discovering 
(1) what are the attributes in the ' certain individuals ' (one 
individual would do as well) on which the ' given attribute ' 
depends, and (2) whether the individual concerning which 
the conclusion is arrived at shares these attributes, is subject 
to the same conditions. A resemblance in respect of other 
attributes than these is nothing to the purpose ; a resem- 
blance to these attributes proves nothing. The question is 
whether the same cause is operative in the two cases ; not 
' whether from the attributes in which Socrates resembles 

1 Book II. chap. iii. sec. 3. individual or individuals resemble the 

2 Mill, Book II. chap. iii. sec. 4. former in certain other attributes; 

3 Book II. chap. iii. sec. 7. ' Certain therefore they resemble them also in the 
individuals have a given attribute ; an given attribute.' 


those men who have heretofore died it is allowable to infer 
that he resembles them also in being mortal/ ! but whether 
Socrates ' has the attributes ' or is subject to the conditions 
on which mortality has been found to depend. When this 
has been settled, there is no further question as to ' whether 
it is allowable to infer ' the mortality of Socrates. Such 
language is a survival of the old notion that inference is 
from what has happened often to what happens always, and 
that the question of induction is whether an event has hap- 
pened often enough to justify this inference. For this Mill 
substitutes the question e whether from the attributes in which 
Socrates resembles those men who have heretofore died it is 
allowable to infer that he resembles them also in being mortal.' 
But from mere resemblance of attributes there is no valid 
inference at all ; and where for such resemblance has been 
substituted an identity between the conditions under which 
Socrates lies and those on which mortality has been shown 
to depend, the inference is over. The process thus described 
by Mill has not then the formal validity of proper syllogism, 
as represented either by the ' dictum de omni et nullo ' or by 
a formula of quantity {top iayaTov sv o\w slvcu to3 fisaqy koX 
tov fLsaov sv oX» to) irpcorcp rj slvai r) fjur) slvac, k.t.X.), nor yet 
is it a process of ' instructive ' reasoning. 

(b) Does it then represent the result of reasoning ? No ; 
in order that it may do so, in the second clause for ' resemble 
the former in certain other attributes ' we must read f are 
identical with the former in respect of other attributes on 
which the given attribute depends.' 2 Such a result of 
reasoning corresponds to Aristotle's * apodeictic syllogism,' 
in which the middle term represents the cause (the formal 
cause or sum of conditions) in virtue of which a certain 
subject (the minor term) undergoes a certain irdOos (the 
major term). E.g. 'the sun with its rays cut off by the 
intervention of the moon (middle term) is eclipsed (major 
term) ; the sun as it now is (minor term) is a sun with its 
rays so cut off; therefore it is eclipsed.' Such a syllogism, 
though no process of demonstration (which has to do with 
showing the dependence of the major on the middle), is the 

1 Book II. chap. iii. sec. 7. a certain other phaenomenon ; these 

2 Better stilly ' Under such and such are phsenomena subject to such coudi- 
conditions certain phgenomona (or indi- tions ; therefore they are accompanied 
viduals as connected groups of phae- (or exhibit) certain other phenomena.' 
nomena) are accompanied by (or exhibit) 


true type of the complete reasoned judgment, which results 
from demonstration. 

Mill's real conception of reasoning, though he speaks as 
if it were represented by his revised formula of syllogism, 
does not appear till the third book. There we find that he 
understands it to consist in the discovery of the causes of 
phenomena, i.e. in a process antecedent to apodeictic syl- 
logism, a process which syllogism cannot represent. ' De- 
duction' indeed he describes as in certain cases entering 
into the discovery, but it is a deduction really quite different 
from syllogism, and which the syllogistic formula will not 



(Mill, Book III.) 

122. Mill gives four definitions of induction 1 ; it is 

(a) * drawing inferences from known cases to unknown 5 ; 

(b) ' affirming of a class a predicate which has been found 
true of some cases belonging to the class ' ; (c) e concluding, 
because some things have a certain property, that other 
things which resemble them have the same property ' ; (d) 
' concluding, because a thing has manifested a property at a 
certain time, that it has and will have that property at other 

(a) is misleading because there is no inference from 
known cases to unknown, except just so far as the unknown 
become known. There is no inference from the mortality 
of other men to that of Socrates, except that which consists 
in coming to know Socrates as we have come to know other 
men in respect of their mortality, i.e. by ascertaining the 
conditions on which mortality depends. Inference consists 
in getting to know Socrates in the same respect, i.e. in 
discovering whether he is subject to the same conditions. 
To (b) the same objection applies. Till you know whether 
the classification is a valid one in respect of the attribute 
affirmed, there is no inference from ' some cases ' to the 
class ; and when you know this, there is no more room for 
inference. So with (c) ; settle the resemblance, as = identity 
of conditions, and the inference is over. As to (d), times as 
such are not in question. Induction has nothing to- do 
either with the times at which observed phenomena have 
been observed, or with other times, except so far as diversity 
of conditions is connected with diversity of times. 

123. The misconception of the nature of induction im- 
plied in the above statements goes along with a misconception 
of the e axiom of the uniformity of nature.' It is regarded 

1 Book III. chap. ii. sec. 5. 

282 LOGIC : MILL. 

as an assumption that things resembling each other in a 
great many points will resemble each other also in others, 
or that what has happened often will happen always, that 
the future will resemble the past. If we ask for the ground 
of such an assumption, we are referred to inductio per 
enumerationem simplicem. A rule which is to enable us to 
dispense with such enumeratio is itself founded on it. Upon 
the strength of a mere enumeration of instances in which 
phenomena have appeared in a uniform relative order, we 
assume from a single instance, in which two phenomena 
have been associated, that they will be in all instances so 
associated. But how do we know that the instances, with 
the examination of which we are always dispensing on the 
strength of the rule, might not be just what would invalidate 
it if they were examined ? If the ground of induction were 
merely an involuntary expectation, it might be accounted 
for in this way. From the constant association of any two 
phenomena we no doubt come to expect the continued 
association of these two, but there is nothing in this to 
produce expectation of continued connexion between others 
which have seldom been presented to us at all, or not in 
nnbroken connexion ; and most of the phenomena with 
which science deals are of the latter sort, of a sort with 
which we only become acquainted at all, or at any rate only 
in unbroken connexion with each other, through ' interro- 
gation of nature,' not through the ordinary course of 
experience. In short, enumeratio simplex, in the sense of 
simple de facto sequence of one feeling on another, could 
simply yield a bundle of expectations of various degrees of 
strength according as the sequence between each series of 
feelings had been more or less frequently repeated and 
unbroken ; nor could the strength of the expectation that 
b will follow d, founded on constant and uniform sequence, 
communicate itself to the expectation that a will follow /, if 
a deficient or varying experience of their connexion left the 
latter expectation weak. Such a bundle of expectations has 
nothing in common with the ground of inductive reasoning, 
as it actually exists. This ground is more fitly expressed 
as the conception of the ' unity of the world ' than as that 
of the ' uniformity of nature, 5 at any rate if the latter is 
supposed to be equivalent to the assumption that the future 
will resemble the past. The future might be exceedingly 


•unlike the past (in the ordinary sense of the words) without 
any violation of the principle of inductive reasoning, rightly 
understood. If the ' likeness ' means that the experiences 
of sensitive beings in the future will be like what they have 
been in the past, there is reason to think otherwise. Present 
experience of this sort is very different from what it was in 
the time of the ichthyosaurus. If it means that different 
experiences of the future will be part of one system with 
the present, the result of conditions that now are, it is 
true ; but to such a system and conditions the distinction of 
past and future does not apply ; they are eternal. On the 
other hand, of that to which the distinction of past and 
future does apply, resemblance cannot be truly predicated. 

124. To make a plausible case for the derivation of the 
principle of induction from an enumeratio simplex of uni- 
formities in the sequence of ' phenomena,' such sequence 
ought to be much more uniform than on the first view it is. 
A certain sight of fire is no doubt uniformly followed by a 
feeling of warmth, &c. &c. ; but, on the other hand, the 
sequences (say) of appearances in the sky seem infinitely 
various. As Mill says, 1 ( the order of nature, as perceived 
at a first glance, presents at every instance a chaos followed 
by another chaos.' Hence when he and others are refuting 
the doctrine that the conception of the ' uniformity of 
nature ' is a priori, which they suppose to mean that every 
man is born with it ready-made, they have no difficulty in 
showing that uneducated men do not believe nature to be 
uniform. They believe in a certain uniformity of nature, 
but in a great deal of wilfulness. How should this be if 
the belief in uniformity is founded on enumeratio simplex, on 
an experience of uniformity which is constantly ' crowding 
in upon us ' ? No doubt it is only upon the first view that 
nature seems a chaos ; that between so many events there 
seems to be no sort of uniform relation. Upon a deeper 
view or ' interrogation ' we find uniformity where there 
seemed chaos. But then it is just this ' interrogation ' 
that has to be accounted for : it is only upon the supposition 
of uniformity that we make the interrogation. How can 
this be, if the supposition is only derived from the observation 
of uniformity, an observation which presupposes the inter- 
rogation ? You cannot come to believe nature to be uniform 

1 Book III. chap. vii. sac. 1. 


till you interrogate her. You cannot interrogate her till you 
believe her to be uniform. 

' If, then, the principle of induction — call it the concep- 
tion of the unity of the world, or what you will — is neither 
derived from observation, as Mill says, nor born ready-made 
with every man, how (it may be asked) do we come by it? ' 
The answer is, that it is implicit in the simplest act of know- 
ledge. (When the human animal begins to know, I do not 
pretend to say.) The unity of the world is the unity of the 
thinking subject. In order to the simplest act of knowledge, 
to that represented by the words c something is,' or ' this is 
here,' a multiplicity of feelings (or, if you like, a feeling 
attended to in successive moments) must become one object 
in virtue of the equal presence of the manifold elements to 
the one subject. The conception on our part of nature as a 
system, of which every part or process is determined by 
relation to all the rest, is merely a development of this 
original determination of our feelings by relation to one 
thinking subject ; and the reality of nature as a system con- 
sists in the relation of its multiplicity to one thinking sub- 
ject, which distinguishes itself from it, but determines it, 
makes it what it is, by this distinction of itself from it. 

125. Thus the definitions which Mill gives of induction 
at best only describe an incident of it, the essence of induc- 
tion being the discovery of the causes of phenomena. What 
is true of a certain phenomenon or sensible event is true of 
all phenomena really the same, i.e. determined by the same 
conditions. This, it will be said, is an identical proposition, 
or another way of putting the principle of contradiction. 
But it is what Mill's formula comes to, if it is to be true at 
all. It is not true, unless 'same' is substituted for ' similar.' 
The whole business of science is to substitute real identity 
(identity of conditions) for mere similarity between pheno- 
mena. The ' resemblance in certain assignable respects ' l 
between the ' all cases ' and the ' particular case ' must be 
identity in respect of the conditions on which the attribute 
predicated depends ; and it is the office of reasoning, whether 
inductive or deductive, to ascertain these. These ascer- 
tained, the work is done. There is no further inference 
from ( some cases ' to ' all cases,' or from c certain times ' to 
f all times.' It is the statement of the conditions of a phe- 

1 Book III. chap. iii. sec. 1. 


nomenon which is the ' general proposition,' in distinction 
alike from the ' singular ' proposition, which merely states 
the occurrence of a phenomenon, and from the ' collective ' 
proposition which summarises any number of such proposi- 
tions. Mill is quite right in saying that it is the business 
of induction to arrive at such truly general propositions. 
But his doctrine about syllogism was that a general propo- 
sition wiis merely a register of a multitude of singular pro- 

126. There are, in Mill, two views of the process by which 
we come to knowledge, which cannot properly be adjusted 
to each other ; one, that it consists in the discovery of 
causes, ' cause' being denned as the sum of the conditions 
of a phenomenon ; the other, that throughout it is the dis- 
covery of resemblances between phenomena, either (a) as 
observation of resemblances between individual phenomena, 
or (b) as abstraction, description, and classification of these, 
or (r) as generalisation, i.e. inference from observed resem- 
blances to unobserved, in which, sharply distinguished from 
(&), Mill considers induction properly to consist. 

Of course, the more completely science is reduced to a 
register of resemblances between phenomena, the less does 
the constitutive action of thought appear in it. Though, in 
truth, it is only for a thinking consciousness that the relation 
of resemblance can exist, yet the existence of such a relation 
for consciousness is so readily confused with the simple suc- 
cession of resembling sensations (which implies no conscious- 
ness of relation), that this may readily be ignored. If all 
the facts, then, which science ascertains consist in ' resem- 
blances between phenomena,' the work of thought in the 
constitution of facts need scarcely come into view. 

It is this work of thought in the constitution of facts 
which Whewell really has to assert as against Mill. But he 
spoils his own case by often writing as if the antithesis be- 
tween ideas and facts were a valid one ; as if the * superin- 
duction of ideas ' upon facts were merely an operation that 
had to be performed ex parte nostra in order to give science. 
Hereupon he is open to the rejoinder that we get our ideas 
from the facts, which is quite true, but is of no avail against 
the true doctrine that it is only the ' colligating ' action of 
thought which constitutes those relations in which the 
c facts ' consist. The true opposition is not between thought 


and fact, but between thought and mere feelings, which, ex- 
cept as related to each other through relation to thought, 
are not facts at all. 1 

127. To return to the acts or processes which Mill dis- 
tinguishes as observation, description (or abstraction), and 
induction (or generalisation) ; the truth is that these are no 
other than stages in one and the same process, by which 
the world becomes to us what it is in itself, a connected 
whole. It is really such a whole in virtue of the presence 
of its manifold to the one eternal thinking subject. To us, 
from the beginning of knowledge, through communication 
in principle of this subject to us as our self-consciousness, it 
is such a whole potentially, i.e. we regard our experience as 
representing a world of which every element is related to 
every other. Only as so regarded is our experience a basis 
of knowledge. It becomes so for us actually, as we come to 
know what the relations between the component parts of 
the world of experience, which from the beginning we pre- 
sume there must be, really are. 

128. Observation, in its simplest form, is the act by 
which we connect manifold feelings in an individual object. 
This connection is not one in the way of resemblance. To 
my simplest apprehension (say) of ( this table ' there go (a) 
feelings which I hold together as immediately successive in 
time, (b) felt objects which I hold together as limiting each 
other in space, and (c) the feelings and felt objects thus 
held together are identified, i.e. regarded as one thing, of 
which the successive feelings are qualities and the mutually 
limiting objects are parts ;. but these relations of space, 
time, and identity, involved in the observation of a single 
object, are none of them resemblance. 

If after an interval I look at the table again, there is no 
doubt a resemblance of the related feelings and felt objects, 
which form the second experience, to those which- formed 
the first ; and in virtue of this similarity of the experiences 
I identify them as representing one object ; ' this is the same 
table that I saw before.' But the relation of identity, 
though it may be thus founded on that of resemblance, is 
quite different from it. We must observe (a) that they are 
not merely resembling experiences that I refer to an identical 
object, but experiences ' contiguous in space and time ' (i.e. 

1 Cf. Whewell, Novum Organum Bcnovaium, p. 116. 


experiences, as Hume would say, related in a way which 
does not depend on anything in the experiences related) ; 
(b) that when it is resembling experiences that are identified, 
the substitution of identity for resemblance is all-important 
as a beginning of knowledge. If the collective representa- 
tion, which we come to denominate as ' this table,' were 
merely regarded as similar, not as representing one thing, 
there would be nothing to be accounted for in the appear- 
ance of an unusual difference between the representations. 
These not being referred to one thing, it would not be a 

129. The ordinary view of observation, which Mill adopts, 
is that groups of sensations having various degrees of resem- 
blance to each other are presented to us ; that then we form 
our idea of the individual object by regarding as one those 
groups which are precisely alike, e.g. (according to the 
language which afterwards comes to be used) the group of 
sensations which I have each time that I look at ' this table ' ; 
that afterwards we combine objects which have less precise 
resemblance ; and so on, as observation passes into abstrac- 
tion, which is supposed to be merely the collection, under a 
name, of points of resemblance between objects which in other 
points more or less differ. Induction, again, is supposed to 
be a discovery of points of resemblance, but not by direct 
observation : it is an inference of unobserved resemblance 
from the observed. 

The fault of this account is that in each stage it ignores 
that determination of objects by relations other than those of 
resemblance which is necessary in order that there may be 
resembling objects to compare. The precisely similar groups 
of sensations, which it supposes us to combine, only are 
similar groups through the intellectual super-induction upon 
mere feelings of relations of time, space, and identity, as 
described above. The sensations which go to make up each 
presentation of ' this table,' are not related to each other 
in the way of resemblance. Nor can that attention to the 
points of resemblance between different individual objects, 
in which ' abstraction ' is supposed to consist, lead to any 
advance in knowledge. Just so far as the science of the 
ancient philosophers consisted in such ' abstraction,' it was 
barren. Take the discovery of early astronomers that 
planets ' revolve in recurring periods' (that each planet so 


revolves as constantly to return to the same position in an 
equal time). The conceptions which this discovery involves 
are none of them got by observation of resemblance. To 
render it possible, there must be (1) the conception 
of motion, of the same body occupying separate spaces 
in successive times (identity, time, space) ; (2) of time as a 
measureable quantity, with something to measure it by ; (3) 
of the planet as revolving ; (4) of a position in space deter- 
mined by relation to the earth, &c. &c. In order to such 
conceptions a long process of determining feelings by rela- 
tions thought of must have gone on, but by relations not in 
the way of resemblance. In each case the conception is 
what Whewell calls a ' colligation of facts,' each fact in turn 
being a colligation of other facts ' nearer to sense,' i.e. imply- 
ing less of the combining action of thought upon the mere 
manifold of feeling. Thus a period of time, a portion of 
space (made up of parts), are facts constituted by intellectual 
colligation. A motion is a further colligation of a period 
of time with a portion of space through the conception of 
body occupying parts of the space in successive times. Nor 
is the conclusion arrived at fitly described as a judgment of 
resemblance. It is the judgment that the periods of time 
occupied by a return of a body to a certain position are 
equal (i.e. qua quantity identical), and the whole value of 
it lies in its being a judgment not of mere resemblance, but 
of equality. From the connection of any set of phenomena 
as merely resembling, no science results : once connect them 
as constituents of a quantity, and we have the beginnings 
of science. Connected as parts of a quantity, they then 
resemble each other in virtue of that relation, but it is 
not in virtue of resemblance that they are so related. All 
things related to each other are similar as a result of that 
relation, but this is quite different from their being related 
in virtue of resemblance. If our knowledge of relations 
results from the observation and abstraction of resemblances, 
the resemblances observed and abstracted cannot be those 
which presuppose relations. Hence our knowledge of relations 
in space and time cannot result from the observation and 
abstraction of resemblance ; for objects cannot be observed 
to resemble each other in these ways unless they have been 
previously known as related in space and time. 

130. All science may rightly be described as progressive 


f colligation of facts ' through superinduction of conceptions, 
if it is understood (a) that ' conception ' means relation, 
which is rightly called c conception ' because it is constituted 
by the combining action of thought upon a manifold ; (b) 
that every fact is constituted by such a superinduction ; 
(c) that thus the colligating conception does not exist in our 
minds before or apart from its existence in fact; and (d) 
that that on which it is superinduced is not the fact as it 
really is, but either (1) feelings on the part of us who feel 
before we understand, or (2) a fact as yet imperfectly con- 
ceived by us, not conceived in the fulness of its relations. 

To describe science as the progressive discovery of the 
conditions of phenomena, conies to the' same thing. The 
word ' phenomenon,' like ' fact,' is ambiguous. Just as 
apart from colligating conceptions there is no fact either 
really or for knowledge, so apart from conditions there is 
no ' phenomenon ' l either really or for knowledge. The 
phenomena which form the data of the most elementary 
knowledge are already conditioned phenomena (conditioned 
by the superinduction of conceptions upon mere appearances 
to sense), or phenomena colligated by mutual relation. 

The first step in knowledge is to connect one appearance 
with another, as forming one object or apparent thing ; to 
identify appearances. This is done by instituting relations 
between them (relations which doubtless really exist, bub 
which for us as sentient are not), and this is to condition them. 
The next step is to connect objects thus formed, in other 
words to condition, by mutual relations, the conditions of the 
first appearances. All knowledge is a continuation of this 
process. To think is to Condition, and to condition is to 
think. The phenomena of which scientific men speak of 
themselves as discovering the conditions are a long way off 
mere appearances to sense ; they are phenomena already 
conditioned by much colligation, highly determinate facts. 
Thus the discovery that air has weight is spoken of as the 
discovery of the conditions, or law, of a phenomenon. 

1 Cf. Deschanel's Natural Philosophy, the same class, we soon perceive that 

p. 4, TV. ' A phaenomenon is a change the various circumstances of their pro- 

that takes place in the condition of a duction have a mutual dependence, so 

body ; the fall of a stone, the flowing that if one of them varies, the others 

of water, the melting of lead, the com- undergo a corresponding variation. The 

bustion of wood, for example, are phae- expression of this connection constitutes 

nomena. When we study the charac- a physical law.' 
teristics which belong to phenomena of 



But in what sense is air a pha3nomenon ? Mill and his 
friends would be prompt to tell us that air is not an ' entity ' ; 
but they do not scruple to call it a phenomenon. Yet in the 
only proper sense of 'entity, 5 as an intelligible object, there is 
much more propriety in calling it ' entity ' than ' phenomenon.' 
It is an intelligible object, but not an appearance to sense, or 
a sensible event. It is an understood relation between the con- 
ditions (themselves relations) of certain sensations. The dis- 
covery that it has weight means that in a certain respect this re- 
lation is id entified with one already known to exist between other 
conditions of phenomena (a relation of bodies to each other), 
and that a quantitative and therefore measurable relation. 

131. It is thus absurd to call the ' law that air has weight ' 
(as Mill seems to do, Book III. chapter iv. § 1) a ' uniformity 
in respect to a single phenomenon.' The air, apparently, is 
the ' single phenomenon ' about which the ' uniformity ' dis- 
covered is that it is always heavy. But air is not a single 
phenomenon, nor even, in any natural sense, a uniformity 
between phenomena ; for uniformity between phenomena 
means a relation between phenomena in the way of resem- 
blance, and the more important relations which constitute 
air are not in the way of resemblance, e.g. motion, and 
production of motion. Nor is the discovery that it has 
weight a discovery of uniformity. If it is, between what is 
the resemblance discovered ? ' Between different cases of 
the phenomenon of air.' But the discovery is that air has 
weight in a single case, when no other conditions than those 
understood by 'air' are present. There is the whole dis- 
covery : there is no further discovery of resemblance between 
that case and all cases. ' The resemblance,' it may be said, 
' is between air and other things that can be weighed.' No 
doubt, if it can be weighed, it resembles other things that 
can be weighed ; but the discovery is of its weight, which is 
not a relation in the way of resemblance, though, of course, 
it constitutes a point of resemblance between all things 
determined by it. The fruitfulness of the discovery lies in 
this, that it connects 'air' (the conditions of phenomena so 
called), or brings it under the same law, with all ponderable 
matter ; renders it a measurable quantity. 

132. The discovery that air has weight is apparently 
what Mill would reckon a proper induction, 1 as distinct (say) 

1 At least he speaks of the law that force which is called its weight,' as an 
air presses upon mercury * with the induction. Book III. chap. iv. sec. 1. 


from Kepler's discovery that planets move in ellipses, which 
is only a description of what has been observed. The one is 
a generalisation, the other only an abstraction; the one 
represents inference, the other only observation ; the one is 
an explanation of phsenomena, the other merely a descrip- 
tion. 1 There is, no doubt, a distinction between such a 
6 description ' of planetary motions as Kepler discovered, and 
such an ' explanation ' of them as Newton discovered, a 
distinction which Whewell puts as that between the laws of 
phsenomena and the laws of. their causes. The question is 
whether Mill gives the right account of it. 

Is it a distinction between a fact seen and a fact in- 
ferred ? No doubt, as Mill says, ' the ellipse was in the 
facts before Kepler recognised it.' That upon which the 
conception of elliptical motion ca,n alone be properly said to 
have been ' superinduced,' consists, not in the facts of Mars' 
positions, but in Kepler's observations of them. The ellip- 
tical motion of Mars may rightly enough be called a 
conception, but in this sense of conception there is no super- 
induction of it. It is always there, constituting the facts. 
That which can alone be said to be superinduced is the 
conception on the part of the astronomer, and that upon 
which this is superinduced is not the facts but the astro- 
nomer's observations. The point is, that in the same sense 
in which Kepler 'saw' the ellipse in the facts, Newton 
' saw ' the law of gravitation in the facts. 

Again, when a conception is said by Mill to be ' abstracted 
from facts ' or ' from phsenomena,' this can only mean that 
it is abstracted from our observations of facts, from the 
facts as they are for the consciousness of the person who is 
supposed to make the abstraction. Otherwise he has nothing 
from which to abstract. But, on the other hand, the 
observations must already be connected and determined by 
the conception (or conceived relation), if it is to be derived 
from them by ' abstraction,' for we cannot abstract what is 
not there to be abstracted. What process, act, or progress 
of thought, then, is represented by this ' abstraction,' by 
which we are supposed to obtain a conception which we 
must already have had in order to the possibility of the 
' abstraction ' ? 

Taking ' facts,' then, in the only sense in which a 

1 Book III. chap. ii. sees. 3 and 4. 

u 2 


conception can with any meaning be said either to be ' super- 
induced upon ' or ' abstracted from ' them, viz. as = our 
observations, the former expression is the more correct of 
the two, because the observations as connected by the con- 
ception take a new character ; a new aspect is superinduced 
upon them. This they must already have, in order that the 
abstraction may be possible, and if they have it, no further 
step is taken in the abstraction, — at any rate, only such as 
consists in giving" abstract expression to (finding a formula 
for) the connecting conception. 

133. Thus, when Mill says, ' Such a conception (that of 
life) can only be abstracted from the phenomena of life 
itself ; from the very facts which it is put in requisition to 
connect,' the answer is that such a statement puts the cart 
before the horse ; that till the phenomena have been con- 
nected by such a conception, they have not the character 
from which it can be abstracted. Doubtless, as Mill says, 
' there is in the facts themselves something of which the 
conception is itself a copy ' (or, more properly, the facts are 
themselves related, through a thought which conceives or 
holds them together, as they come to be for us) ; but from 
the facts themselves, as distinct from our observations, we 
can make no abstraction. The business of science is to 
connect our observations (the facts as they are, or are con- 
stantly coming to be, for our consciousness) by the con- 
ception by which the facts themselves are connected ; to 
reproduce in us this conception. Such reproduction is only 
possible because, the thinking subject which is the unity of 
the world being in principle present in us as our reason, 
our observations, or the facts as they are for our conscious- 
ness, are already potentially what they are in themselves. 
We are constrained to seek to think them or hold them 
together as one, and the only way in which this can be done 
is by connecting them as they are really connected. Just 
in so far as we fail so to connect them, the facts of our 
observation (or the facts as they are for our consciousness) 
are a contradiction to that unity which, because in itself, 
thought must seek to find in the world. Just so far as we 
succeed in so connecting them, the facts of observation (or 
the facts as they are for our consciousness) become the real 
facts. Thus it is the ' phenomena of life ' that bring us to 
the true conception of life, not in the sense that we abstract 


it from them, but in the sense that the facts of life, as 
observed or as they are for our consciousness, are a contra- 
diction, a perplexity, a baffling manifold, in which thought 
cannot rest till they are connected for us as they are really 
connected, till these phenomena become the realities. 

134. Mill's distinction, then, between Kepler's discovery 
and Newton's, so far as it depends on the view that the 
former is an ' abstraction,' of a kind with that of the old logic, 
breaks down. The conception of the relation between the 
positions of Mars as points in an ellipse is not abstracted 
from Kepler's observations of them, for, till they have been 
determined by this conception, they have not the common 
characters from which it could be abstracted. They have 
first to be determined by the conception of their mutual 
relation as points in an ellipse (which is not a relation in 
the way of resemblance, though as all alike determined by 
it they resemble each other), before the abstraction can be 
made. Hence the fallacy of Mill's statement, 1 * The mental 
operation which extracts from a number of detached obser- 
vations certain general characters in which the observed phe- 
nomena resemble one another, or resemble other known facts, 
is what Bacon, Locke, and most subsequent metaphysicians 
have understood by the word abstraction,' Cf. III. ii, § 4. 
1 The assertion that the planets move in ellipses was but 
a mode of representing observed facts ; it was but a colli- 
gation ; while the assertion that they are drawn, or tend, 
towards the sun, was the statement of a new fact, inferred 
by induction.' Write ' newly known fact ' for ' new fact ' in 
the second clause, and it becomes equally applicable to 
Kepler's discovery. 

As little as Kepler's discovery and others which Mill 
would class with it are < abstractions ' in the ordinary sense, 
is what he reckons as < induction proper,' such as Newton's 
discovery, a ' generalisation ' in the ordinary sense, as an 
inference from some cases to all, from known cases to 
unknown. Mill says, 2 ' The universe, so far as known to us, 
is so constituted, that whatever is true in one case is true in 
all cases of a certain description ; the only difficulty is to 
find what description.' The business of induction, then, it 
appears, is to find a true and adequate description of the 
single case, to find the sum of its conditions. This done 


1 Book III. chap. ii. sec. 5. ■'••■»•« /^ # c hap. iii. see. 1. 


everything is done. Till it is done, generalisation to all like 
cases may be wholly false, for apparent likeness is com- 
patible with entire diversity of conditions ; when it is done, 
generalisation is purposeless, for the cases to which we 
extend the generalisation are, if it is to be valid, the same 
in respect of their conditions with that which is the basis of 
the generalisation. The only sense in which generalisation, 
as a process from the observed to the unobserved, has a 
place in knowledge, is as an anticipation of what results 
from the combination of conditions already known in their 
separate, but not in their joint, action, i.e. as the process 
which Mill calls ' deduction.' 

135. The distinction, then, between observation and in- 
ference breaks down, as does that between abstraction or 
description and generalisation, as an account of the differ- 
ence between the discoveries of Kepler and Newton. There 
is inference in the simplest observation, if everything is 
inference which goes beyond sense, for already in such 
observation there is a determination of phenomena in the 
strict sense (as = appearances) by conditions consisting in 
understood relations, and there is no more inference than 
this in any induction. The true difference between such a 
c description ' as that of Kepler and such an ' induction ' as 
that of Newton (which, according to Mill's own showing, 1 is 
only a more complete description), lies in the range of 
the ' colligation ' which they severally imply. The discovery 
of Kepler enables us to ' colligate - the observed position of 
the planets ; that of Newton (I believe) all motions of bodies 
whatever. That of Kepler is merely a discovery of the 
mode in which the positions of planets condition each other.; 
that of Newton is of the dependence of these mutually con- 
ditioning positions upon a condition common to all matter. 

The wider the colligation, the greater the range of facts 
unified by a conceived relation, the less becomes the possi-r 
bility of the relation as conceived by us being other than 
the relation as it really is, or as it is according to the true 
conception. Whewell says that often several different con- 
ceptions will serve equally well to colligate the same set of 
observed facts. To which Mill rightly replies 2 that though 
these several conceptions may serve equally well as a descrip- 

1 [P. 337. See passage quoted .on the preceding page from Mill, III. iii. 1.] 

2 Book III. chap. ii. sec. 4, note. 


tion of the facts, they will not as an explanation of them. 
The rationale of this is, that a set of facts, though described 
by their colligation with each other (the statement of their 
relations as conditioning each other), is only explained by 
colligation with other sets of facts. In this respect Mill's 
account of the difference between the discoveries of Kepler 
and Newton as one between description and explanation 
may be accepted. Of course a description which does not 
admit of explanation is not really a true description (though 
Whewell sometimes writes as if it might be, just as he some- 
times writes as if the conception by which facts are united, 
instead of being a relation belonging to the facts, merely 
existed ex parte nostra) . If our conception of the facts were 
the conception (relation) which really connects them, on the 
principle that the world is one it would admit of connection 
with the conception by which other facts are connected under 
some common conception, i.e. it would admit of explanation. 
Thus the truth is not that two descriptions may be equally 
true, two explanations not, but that the more complete colli- 
gation is the test of the truth of the less complete. Of two 
descriptions the untruth of one may be settled by the im- 
possibility of explaining it, i.e. of colligating it with other 
groups of facts. But the explanation by which the true 
description is colligated with other facts is only a wider 
description, 1 which may in turn be found inadequate because 
not admitting of more complete colligation. 

1 Book III. chap. ii. sec. 4, note. 



(Mill, Book III., Chapter V.) 

136. It is commonly supposed that there are two views 
about causation, between which our choice lies; one that 
it is simply a relation of uniform sequence between one 
phenomenon and another or others (to ' uniform ' some 
would add * unconditional/ without supposing that it makes 
any difference) ; the other that it consists in a * mysterious 
tie ' between one phenomenon and another, or in a power 
exercised either by a natural agent, in virtue of which what 
is called its effect follows, or by something external to both 
the nominal cause and the effect, which determines the 
sequence of one upon the other. In fact, however, though 
the doctrine of the * mysterious tie ' is always being ascribed 
by English e experimentalists ' to people whom they call 
' transcendentalists,' they never condescend to tell us what 
' transcendentalist ' in particular holds the doctrine. 

Another common notion is that there are two different 
questions, one as to the nature of causation (the relation of 
cause and effect) itself, another as to the origin of our idea 
of causation, Thus there are people who hold that the 
origin of our idea of causation lies in our consciousness of 
volition or voluntary effort ; in our experience of ability to 
move muscles, and through them other things, upon a pre- 
ference. But people who hold this do not (at any rate 
always) hold that the relation of cause and effect in nature 
implies such volition on the part of the cause or on the part of 
an omnipresent agent. There are those, again, who hold that 
the constancy in the sequence of certain feelings upon others 
is what gives us the idea of the relation of cause and effect ; 
but who would not admit that this relation itself consists in 
such constant sequence or in the habit of expectation pro- 
duced by it ; who regard it, on the contrary, as belonging to 
an objective order of nature on which the sequence of our 
ideas may depend, but which is not interchangeable with it. 


If we think the matter out, however, we shall find that 
the question of what the relation of cause and effect is in 
itself is identical with the question of the source of our idea of 
it. The conception ex parte nostra is merely a reproduction, 
more or less complete, of the conception (or relation) as it 
really exists. 

137. The characteristic of Hume's doctrine as stated by 
him is, that, according to it, the relation of cause and effect 
is the succession of our impressions and ideas, or an ' im- 
pression of reflection ' consisting in the habit of expectation 
derived from this, not an * objective order ' on which the 
succession of our ideas depends. The sight of flame has 
constantly been followed by the feeling of heat ; hence, upon 
seeing flame, the idea of heat presents itself involuntarily 
with great liveliness. In this liveliness and involuntariness 
with which the feeling of heat is expected upon flame being 
seen consists, not merely our idea of heat as the effect of 
flame, but the connexion itself in the way of cause and effect 
between flame and heat. 1 This, according to Hume, is the 
account of the connexion between every particular effect and 
every particular cause. Of any general law of causation 
Hume (according to his own showing) knows nothing. The 
only account that he could consistently give of such a law 
would be that it is the sum of all particular habits of expec- 
tation, of the kind just described. To take another illustra- 
tion (which Hume uses in the * Essays '). One billiard ball 
strikes another, and this other moves. There is an imme- 
diate sequence in time between the motion of the latter ball 
and the impact of the former ; but there is also an immediate 
sequence between, e.g. my touching this table and seeing 
the clock, yet I do not reckon my seeing the clock an effect 
of my touching the table. In what lies the difference be- 
tween the two cases ? The fact that the succession in the 
one case has been observed constantly, in the other perhaps 
never before, makes no difference to the succession or to the 
events between which the succession obtains. What it does 
make a difference to is my habit of expectation. If I shut 
my eyes at the moment of the impact of one ball on the 
other, I should ' believe in ' (have the liveliest possible idea 
of) an ensuing motion of the other ball. But if I shut my 
eyes or failed to turn my head at the moment when I touch 

1 See vol. i. General Introduction to Hume, sec. 284 if., especially sec. 292. 


this table, no idea of the clock would present itself to me. 
And in this subjective difference (according to Hume) lies 
all the difference between a sequence which is, and one which 
is not, a relation of cause and effect. It is thus not a true 
account of the matter to say that, according to Hume, the 
connexion between any cause and any effect consists in the 
invariable, or even the usual, sequence of the latter upon the 
former. Hume was acute enough to see that the mere re- 
petition of the succession of events makes no difference to it, 
as a succession of events ; that one sequence in time is 
exactly like every other, the sequence of the motion of one 
billiard ball on that of another an exactly similar sequence 
to that of the sight of the clock upon the touch of the table. 
What ordinary people really mean when they speak of the 
relation of cause and effect as invariable sequence is, that 
the invariability of a sequence is a test that the sequence of 
events is much more than a mere sequence of events, that it 
is a sequence so determined by the system of nature as that one 
event cannot occur without the other ; and this implies neither 
! power ' nor a ' mysterious tie,' on the contrary, it is strictly 
and solely intelligible. It is such determination, of which 
invariability is merely a test, which really distinguishes the 
relation of cause and effect from other sequence. But deter- 
mination by a system of nature is neither an 'impression' 
nor an * idea,' nor a succession of impressions and ideas, nor 
can it be -represented by them ; hence, logically, Hume can 
know nothing of it, and he is so far logical that though he 
cannot help using language which implies it, he dispenses 
with it as far as he can, — will not deliberately avail himself 
of it in his system. Thus, since some difference has to be 
found between antecedence in the way of causation and 
other antecedence, and since it cannot be found in any 
difference which repetition makes to the antecedence as such 
(one antecedence being like every other), it remains for 
Hume to find it, as he does, in the feeling of expectation on 
our part with which the impression of the antecedent is 

This doctrine could scarcely be adopted by any one who 
understood what it amounted to. It would follow from it 
that causation admitted of degrees, as the habit of expecta- 
tion does. It is sometimes thought that Hume only needed 
to have been aware of the doctrine of 'hereditary trans 


mission ' &c. in order to have stated his doctrine in an 
unassailable way. The only difference that this could have 
made to his doctrine would have been that, whereas in its 
present form it would seem to follow from it that causation 
was a relation which came gradually into being with each 
individual's habits of expectation, as reformed by the doc- 
trine of hereditary transmission it would have led to the 
conclusion that causation was a relation which was gra- 
dually coming into being with the growing experience of the 
human race. 

Undoubtedly, according to Hume's doctrine, the relation 
between day and night would be one of cause and effect, as 
much as any relation of antecedent and sequent can be. 
The expectation of night as to follow, which is incidental to 
the experience of the day, is as strong as any expectation 
can be. 

138. Scientific men often suppose themselves to adopt 
Hume's doctrine of cause and effect, but in fact they only 
do so with two alterations, which make it not only quite a 
different doctrine, but one which could not have been 
logically arrived at from Hume's premisses. They do not 
hold that the difference between a sequence of one event on 
another which is, and one which is not, a sequence of effect 
on cause, lies in the difference that a habit of expectation 
accompanies the one and not the other. They do hold that 
the reason why invariability is essential to sequence in the 
way of effect is, not because without invariability the habit of 
expectation would not be formed, but because a sequence in 
the way of effect is one determined by a system of nature, 
so that from any cause only one effect can really follow. 

To revert to the distinction between the sequence of the 
motion of one billiard ball on that of another and the 
sequence of my sight of that clock on my touch of this 
table, no scientific man would admit that the difference lay 
in the fact that the former sequence had constantly been 
repeated. They regard a sequence between events as one of 
causation on the first time of its occurrence. In the words 
of Mill, 1 'Between the phenomena which exist at any 
instant, and the phenomena which exist at the succeeding 
instant, there is an invariable order of succession.' Thus 
the state of the earth at any time is the effect of its state as 

1 Book III. chap. v. sec. 2. 


determined by its relation to the rest of the universe at the 
immediately preceding time, but neither state has ever 
occurred before or will ever occur again. ' Invariable,' then, 
here cannot mean ' invariably repeated,' but only ' which 
could not be other than it is.' 

The residuum of itself, in short, which Hume's doctrine 
has left in ordinary scientific men is, (a) that in the last 
resort there is nothing in any phenomenon to account for 
the uniform sequence of another upon it; (b) that our 
notion of, or belief in, causation, as distinct from that 
relation itself, results from repeated experience (transmitted 
from generation to generation). 

139. (a) We find, it is said, that as a matter of fact 
phenomenon a generally follows phenomenon b, but that 
there are exceptions. We seek to explain these in order to 
find the true cause of a, and find that it only follows b 
when b itself follows certain other phenomena, and again 
that b only follows these when these occur in a certain 
sequence, and so on. We never can get beyond the fact 
that one phenomenon always follows another, or only 
follows that other when that other follows something else ; 
we never can give a reason why it should be so. When we 
give a Blotl for a natural on we are merely stating anoth er 
oti. When we are said to investigate the nature of any 
phenomenon in order to explain the sequence of another 
upon it, we are merely ascertaining certain de facto sequences 
which constitute the nature of a complex phenomenon, and 
for none of these can we in the last resort give a reason. 1 

In this, it is said, lies the difference between ' a cause * 
and ' a reason,' in the confusion between which the great 
error of ancient philosophy lay (though Aristotle distin- 
guishes alriov yvoxTscos and alriov ysvscrscos.) From a reason 
you can infer the consequent, from a cause you cannot 
infer the effect. If one phenomenon has always been 
found to be followed by another, you will expect upon its 
occurrence that the usual sequent will follow, and only in 
that sense do you infer effect from cause. 

So Hume said, ' No idea or object considered in itself can 
give a reason for drawing a conclusion beyond it.' In 
reasoning, according to him, you do not ' go beyond ' a 
given idea, but merely break it up. But such reasoning 

1 See Mill, Book III. chap. xii. sec. 6. 


has nothing to do with ' matters of fact ? ; it merely deals 
with 6 nominal essences.' No ' matter of fact,' according to 
him, could be ' proven ' or \ inferred ' or ' deduced ' at all. 
His modern followers are not so strict. If matters of fact 
cannot be proven, inferred, or deduced, there is an end of 
inductive logic. Yet the exponents of inductive logic com- 
monly accept the antithesis between ' reason ' and ' cause,' 
between the order of facts and the order of thought, and 
the doctrine that the relation of cause and effect is only a 
relation ofuniform sequence between phenomena, of which 
the denial of the possibility of j proving ' or ' inferring ' 
facts is the corollary. * Can you,' it may be said, ' by any 
amount of reasoning make known a fact which was not 
known before P No. But causes and effects are alike 
matters of fact. Then no cause can be discovered from an 
effect or any effect from a cause by reasoning, nor can there 
be anything in any particular cause to account for its 
particular effect, nor anything in any effect to imply its 
cause ; there can be no reason why one should precede or 
follow the other.' This is the Humian view, which English 
empiricists have hitherto supposed themselves to adopt. 

140. It is quite true that by 'mere reasoning,' if that 
means syllogism (the evolution of the content of a given 
idea or, more properly, of the meaning of a name), no fact, 
not already known, can be known ; but it is equally true 
that a ' mere fact,' a fact apart from relations which are not 
sensible, would be no fact, would have no nature, would not 
admit of anything being known or said about it. No 
reasoning can yield new experience in the way of feeling, 
but new experience in the way of feeling merely or by itself 
is no intelligible fact, no addition to knowledge. 

Thus, to Hume's maxim, quoted above, we reply that 
6 no idea or object ' can be ' considered in itself ' ; and just 
because it cannot, every ' idea and object ' (every experience 
in the way of feeling, every object to which we refer such 
experience) compels a conclusion beyond it. The ' minimum 
intelligibile ' in the way of feeling (the only experience which 
amounts to a knowable fact) is a feeling related to another 
as a changed appearance or affection of something of which 
the other was an appearance or affection, whether that 
' something ' be regarded as a feeling subject or a felt 
object. The conception of this something developes, as 


everything is found to be relative to another and to derive 
all that it is or has from that relation, till the ' something ' 
becomes * nature ' (of which Lewes has at last discovered 
that to say that it is uniform is an identical proposition) 
which remains the same in all its changes. 

This minimum intelligibile, however, is still unintelligible. 
The identical changes. Without identity there is no change, 
yet change contradicts identity. To overcome the contra- 
diction the change must be accounted for. At first the 
changes seem chaotic. The first thing (a) to be done for 
satisfying that demand for unity, which is the ground of 
our intelligence and at the same time identical in principle 
with the unity (the one subject) of the world, is to make 
out what follows what ; the next thing (b) is to account for 
the ascertained uniformity by progressive discovery of its 
conditions, a discovery which at once further ascertains the 
uniformity and connects it with other uniformities of change. 
(No doubt natural philosophy has been held back by the 
hurry to get to (6) without sufficient care in (a), and by 
acquiescence as regards (b) in mere verbal explanations or 
appeals to supposed agencies extra-natural, assumed ex 
analogia hominis.) 

Thus, as no real object is a separate object, so no known 
object, in being known, can be considered by itself. * What 
is beyond it ' constitutes both its reality and its knowability. 
Considered in relation to the knowability of the object, 
' what is beyond it ' is the reason of which the object as 
known is the consequence ; considered in relation to its 
reality, ' what is beyond it ' is the cause of which the object 
as it exists is the effect. Or, conversely, since the object, so 
far as known, determines that beyond it which, in knowing 
it, I am coming to know, and in its existence determines 
that beyond it which its existence implies, this ' beyond it,' 
considered in relation to its knowability, is the consequence 
of which it is the reason, considered in relation to its exist- 
ence, is the effect of which it is the cause. 

141. In short, the absolute antithesis between the rela- 
tion of reason and consequence and that of cause and effect 
is part of the false antithesis between thought and reality, 
which goes along with the reduction of reality to mere 
individuals, whether e things ' or •' events.' Once apprehend 
(what is implied in all the teachings of science) that there 


are no isolations or separations in nature, that 'individuals ' 
are mere logical fictions (that, strictly, individuality is a 
logical category which has no reality except in correlation 
with all other categories), that no event happens which is 
not determined by, and does not contribute to determine, 
the whole system of nature ; once apprehend this, and the 
notion that the relation of cause and effect is fitly described 
as that of an invariably preceding to an invariably following 
event must be given up. Any effect in its reality = its 
cause. What is the cause of water ? Does this mean the 
cause of an event consisting in the formation of some 
water ? If so, we must answer that the cause of water is 
the combination of hydrogen and oxygen in certain pro- 
portions. But this combination, which is rightly said to be 
the cause of the event consisting in that formation, = that 
event. There is no antecedence in time of cause to effect. If 
by ' water,' however, is meant a composite chemical substance 
having certain properties (as in Mill, Book III. chapter x. § 4), 
then we must say that the cause of water lies in oxygen and 
hydrogen, as combined in certain proportions, but still there 
is no antecedence of cause to effect. The cause = the effect, 
and the effect = the cause. The view of cause, in short, as 
an event uniformly preceding another event, is incompatible 
with the definition of it as the sum of the conditions of a 

142. ' But,' it may be replied, * though the cause be not 
a preceding event, but the sum of conditions, these conditions 
are all events. Even the ' permanent causes ' which Mill 
admits l may be reduced to groups of events, sequent or 
contemporary, as may the c objects ' which, according to him, 
'enter as causes into the sequences called states of those 
objects.' 2 Thus, though the relation of cause and effect 
may not be resoluble into uniform sequence, it is resoluble 
into a multitude of sequences and coincidences taken to- 
gether ; and our ultimate analysis cannot get beyond the 
mere fact, for which no reason can be given, that certain 
events are simultaneous with, certain events successive npon, 
certain other events ; and our laws of nature are merely 
summary statements of such simultaneities and successions. 
Thus the appearance of the phamomenon ' water ' is coin- 
cident with, not sequent upon, the combination of oxygen 
1 Book III. chap. v. sec. 7. 2 Ibid. sec. 4. 


and hydrogen ; we know no reason why their combination 
should constitute such an appearance, and though the 
formation of water is not sequent upon the combination, 
the characteristic of the formation — that which we mean by 
distinguishing it as water — is that certain other phenomena 
will follow. Thus the ' essence ' of water consists (a) in its 
formation upon a certain coincidence, and (b) in the sequence 
of certain events upon that formation. Facts of simultaneity 
and sequence make up its nature, as a wider range of such 
facts make up all nature.' 

143. First let us be clear what coincidence or simul- 
taneity means. It is a designation of events, not, indeed, 
sequent on each other, but each sequent upon one and the 
same event in that particular series of recurrent events by 
which we measure time (the diurnal motion of the sun). 
A relation of simultaneity, then, just as a relation of se- 
quence, (a) implies that the related objects are in time, and 
(b) is only possible for (in relation to) a subject not itself in 
time, but equally present to the succession of times ante- 
cedent, to the time at which the simultaneous events occur, 
and to that time itself. Thus, if all reality were reducible 
to a multitude of connected successions and simultaneities 
(we must add 'connected' if such an account of reality is to 
have any appearance of corresponding with science), there 
would still be implied a single subject to which all these 
were relative. And it would still be misleading to speak (as 
Mill does ! ) of nature as made up of separate uniformities in 
respect of simultaneity and succession, since this conveys 
the notion that each uniformity is independent of all the 
rest ; which is to reduce the world to chaos. That which 
gives its character to any sequence or simultaneity (that 
character which science seeks to ascertain) is not the number 
of instances in which the sequence or simultaneity has 
occurred, though that is what we seem to imply when we make 
' uniformity ' the differentia of the sequence or simultaneity ; 
(the simultaneity between the appearance of life inaparticular 
part of the earth and the attainment of certain conditions in 
respect of temperature and otherwise could but occur once, 
yet it is not the less a determined simultaneity, which could 
not have been other than it was) ; it is its relation to the 
other simultaneities and successions which, if it be so, form 
the system of nature. Now this relation of all simul- 

1 Book III. chap. iv. sec. 1. 


taneities and successions to each other so that one could not 
be without the rest, which is alone the adequate cause of 
any phenomenon, is not itself simultaneous with, or suc- 
cessive upon, anything. It is not an event — not in time — 
and the designations of simultaneity and sequence are only 
applicable to events. 

Thus, admitting it to be true that every ' fact ' is a 
simultaneity or succession of events or appearances, and 
that the system of nature is the sum of all such 'facts,' 
still, inasmuch as the system of nature can only = the sum 
of such facts as determined by relation of all to each other, 
which is not a relation in the way of time, that which gives 
it its character (makes it what it is) is not simultaneity or 
succession or both together, but a unity which is properly 
eternal (not an event or any number of events), and to 
which designations appropriate to what is in time are w T hollj r 
inapplicable. In like manner the relation of cause and 
effect is not a relation in time, not one to which either 
' simultaneity ' or ' succession ' are applicable. Cause is 
defined as the 'unconditional antecedent,' but the two terms 
of the definition are incompatible with each other. You 
can find no unconditional antecedent short of the whole 
system of nature (for there are no events of which you can 
say that one must follow the other though all the con- 
ditions of the universe were changed), and to that the 
term ' antecedence ' has no proper application. 

144. Hume held that there could properly be no cer- 
tainty, but only probability, in regard to a relation of cause 
and effect. With him, as we have seen, the relation itself 
consists in a lively belief, as distinct from that knowledge of 
which the only possible object is ' relations of ideas ' (i.e. 
the relation of one idea to another as containing or con- 
tained in it, and mathematical relations, according to the 
doctrine of the Essays, which is a return to Locke). Such 
a lively belief admits of various degrees, according to the 
amount and uniformity of the experience on which it is 
founded, but it never reaches certainty. With Mill this 
distinction between belief and knowledge has disappeared. 
The relation of cause and effect is an objective relation, the 
correct copy of which in our minds constitutes certain 
knowledge, though there may be always some doubt whether 
we have attained such a copy. At the same time the 

VOL. II. x 


relation is supposed to consist in the sequence of one event 
on another, so that our certainty in regard to it consists 
in a conviction that the sequence will always continue, 
' always ' being taken to mean ' unconditionally.' This 
conviction in regard to any particular sequence is supposed 
to be based (a) on the belief that on every event some other 
(whatever that other may be) without exception follows, a 
belief which simply results from our finding in ordinary 
experience that it always is so ; (b) on such an examination, 
governed by that belief, of the complex antecedents of the 
particular event, as ascertains which of them may be absent 
without the event ceasing to happen. Those which cannot 
be so absent are unconditional antecedents. 

On this it is to be remarked as regards (a), that the 
experience of the constant sequence of event c on event d 
may doubtless lead to the strongest expectation of one on 
occasion of the other. But how should that cause the 
belief that e will follow/, when there has been no constant 
experience of it, or lead to an interrogation of nature in 
order to explain the apparent irregularity in the sequence 
of e on /? In fact all the attempts to explain ' belief in 
uniformity,' as resulting from the passive experience of 
constancy in the sequence of events, presuppose some 
rudimentary conception of nature. Without this, such 
experience could only yield . a bundle of expectations, of 
which one might indefinitely strengthen or weaken another, 
but of which none could afford any explanation of another. 
With this rudimentary conception (of which the true account 
is that it is the presence in us as our self-consciousness of 
the single subject which is presupposed in the possibility of 
a nature), the several constancies and inconstancies become 
constantly more and more explanatory of each other. As 
to (b), if the supposition of an unconditional sequence could 
be arrived at in the wa} r suggested, it would be an un- 
warranted and misleading one. 




Note of the Editor. 

The lectures from which the following extract is taken were delivered 
in the beginning of 1 879, in continuation of the course in which the dis- 
cussion of Kant's moral theory occurred. The portions here printed are 
those which were not embodied, at any rate in the same form, in the Pro- 
legomena to Ethics. 

x 2 





1. Since in all willing a man is his own object, the will 
[ is always free. Or, more properly, a man in willing is 
necessarily free, since willing constitutes freedom, 1 and 'free 
will ' is the pleonasm ■ free freedom. 5 But while it is 
important , to insist upon this, it is also to be remembered 
that the nature of the freedom really differs — the freedom 
means quite different things — according to the nature of the 
object which the man makes his own, or with which be 
identifies himself. It is one thing when the object in 
which, self-satisfaction is sought is such as to prevent that 
self-satisfaction being found, because interfering with the 
realisation of the seeker's possibilities or his progress 
towards perfection : it is another thing when it contributes 
to this end. In the former case the man is a free agent in 

/the act, because through his identification of himself with 
a certain desired object — through his adoption of it as his 
good — he makes the motive which determines the act, and 
is accordingly conscious of himself as its author. But in 
another sense he is not free, because the objects to which 
his actions are directed are objects in which, according to 
the law of his being, satisfaction of himself is not to be 
found. His will to arrive at self-satisfaction not being 
adjusted to the law which determines where this self- 
satisfaction is to be found, he may be considered in the 
condition of a bondsman who is carrying out the will of 
another, not his own. From this bondage he emerges into 
real freedom, not by overcoming the law of his being, not 

1 In that sense in which 'freedom' expresses a state of the soul, as distinct 
from a civil relation. 


by getting the better of its necessity,-- every fancied effort 
to do so is but a new exhibition of its necessity, — but by 
making its fulfilment the object of his will ; by seeking the 
satisfaction of himself in objects in which he believes it 
should be found, and seeking it in them because he believes 
it should be found in them. For the objects so sought, 
however various otherwise, have the common characteristic 
that, because they are sought in such a spirit, in them self- 
satisfaction is to be found ; not the satisfaction of this or 
that desire, or of each particular desire, but that satisfaction, 
otherwise called peace or blessedness, which consists in the 
whole man having found his object ; which indeed we never 
experience in its fulness, which we only approach to fall 
away from it again, but of which we know enough to be 
sure that we only fail to attain it because we fail to seek it 
in the fulfilment of the law of our being, because we have 
not brought ourselves to ' gladly do and suffer what we must.' 

To the above statement several objections may be made. 
They will chiefly turn on two points ; (a) the use made of the 
term ' freedom ' ; (b) the view that a man is subject to a 
law of his being, in virtue of which he at once seeks self- 
satisfaction, and is prevented from finding it in the objects 
which he actually desires, and in which he ordinarily seeks it. 

2. As to the sense given to ' freedom,' it must of course be 
admitted that every usage of the term to express anything but 
a social and political relation of one man to others involves 
a metaphor. Even in the original application its sense is by 
no means fixed. It alwa} 7 s implies indeed some exemption 
from compulsion by others, but the extent and conditions 
of this exemption, as enjoyed by the ' freeman ' in different 
states of society, are very various. As soon as the term 
' freedom ' comes to be applied to anything else than an esta- 
blished relation between a man and other men, its sense 
fluctuates much more. Reflecting on their consciousness, on 
their ' inner life' (i.e. their life as viewed from within), men 
apply to it the terms with . which they are familiar as 
expressing their relations to each other. In virtue of that 
power of self-distinction and self- objectifi cation, which he 
expresses whenever he says ' I,' a man can set over against 
himself his whole nature or any of its elements, and apply to 
the relation thus established in thought a term borrowed 
from relations of outward life. Hence, as in Plato, the terms 


1 freedom * and ( bondage ' may be used to express a relation 
between the man on the one side, as distinguishing himself 
from all impulses that do not tend to his true good, and 
those impulses on the other. He is a 6 slave ' when they are 
masters of him, ' free ' when master of them. The metaphor 
in this form was made further use of by the Stoics, and 
carried on into the doctrines of the Christian Church. Since 
there is no kind of impulse or interest which a man cannot 
so distinguish from himself as to present it as an alien 
power, of which the influence on him is bondage, the parti- 
cular application of the metaphor is quite arbitrary. It may 
come to be thought that the only freedom is to be found in 
a life of absolute detachment from all interests; a life in 
which the pure ego converses solely with itself or with a God, 
who is the same abstraction under another name. This is a 
view into which both saints and philosophers have been apt 
to fall. It means practically, so far as it means anything, 
absorption in some one interest with which the man iden- 
tifies himself in exclusion of all other interests, which he 
sets over against himself as an influence to be kept aloof. 

With St. Paul the application of the metaphor has a 
special character of its own. With him ' freedom ' is specially 
freedom from the law, from ordinances, from the fear which 
these inspire, — a freedom which is attained through the com- 
munication of what he calls the ' spirit of adoption ' or ' son- 
ship.' The law, merely as law or as an external command, is 
a source of bondage in a double sense. Presenting to man a 
command which yet it does not give him power to obey, it 
destroys the freedom of the life in which he does what he 
likes without recognising any reason why he should not (the 
state of which St. Paul says ' I was alive without the law 
once ') ; it thus puts him in bondage to fear, and at the same 
time, exciting a wish for obedience to itself which other 
desires (<f>povrnia aapKos) prevent from being accomplished, it 
makes the man feel the bondage of the flesh. * What I will, 
that I do not ' ; there is a power, the flesh, of which I am the 
slave, and which prevents me from performing my will to 
obey the law. Freedom (also called ' peace,' and ' reconcilia- 
tion ') comes when the spirit expressed in the law (for the 
law is itself ' spiritual ' according to St. Paul ; the ' flesh ' 
through which it is weak is mine, not the law's) becomes the 
principle of action in the man. To the man thus delivered, 


as St. Paul conceives him, we might almost apply phraseology 
like Kant's. ' He is free because conscious of himself as the 
author of the law which he obeys.' He is no longer a ser- 
vant, but a son. He is conscious of union with God, whose 
will as an external law he before sought in vain to obey, 
but whose ' righteousness is fulfilled ' in him now that he 
* walks after the spirit.' What was before ' a law of sin and 
death ' is now a ' law of the spirit of life.' (See Epistle to 
the Romans, viii.) 

3. But though there is a point of connection between St. 
Paul's conception of freedom and bondage and that of Kant, 
which renders the above phrase applicable in a certain sense 
to the ' spiritual man ' of St. Paul, yet the two conceptions 
are very different. Moral bondage with Kant, as with Plato 
and the Stoics, is bondage to the flesh. The heteronomy of 
the will is its submission to the impulse of pleasure- seeking, 
as that of which man is not in respect of his reason the 
author, but which belongs to him as a merely natural being. 
A state of bondage to law, as such, he does not contemplate. 
It might even be urged that Kant's ' freedom' or 'autonomy' of 
the will, in the only sense in which he supposed it attainable 
by man, is very much like the state described by St. Paul as 
that from which the communication of the spirit brings de- 
liverance, — the state in which 6 1 delight in the law of God after 
the inward man, but find another law in my members warring 
with the law of my reason and bringing me into captivity to 
the law of sin in my members.' For Kant seems to hold that 
the will is actually ' autonomous,' i.e. determined by pure 
consciousness of what should be, only in rare acts of the best 
man. He argues rather for our being conscious of the pos- 
sibility of such determination, as evidence of an ideal of what 
the good will is, than for the fact that anyone is actually so 
determined. And every determination of the will that does not 
proceed from pure consciousness of what should be he ascribes 
to the pleasure- seeking which belongs to man merely as a 
' Natur-wesen,' or as St. Paul might say * to the law of sin 
in his members.' What, it may be asked, is such ' freedom,' 
or rather such consciousness of the possibility of freedom, 
worth ? May we not apply to it St, Paul's words, ' By the 
law is the knowledge of sin ' ? The practical result to the 
individual of that consciousness of the possibility of freedom 
which is all that the autonomy of will, as really attainable by 


man, according to Kant's view, amounts to, is to make him 
aware of the heteronomy of his will, of its bondage to motives 
of which reason is not the author. 

4. This is an objection which many of Kant's statements 
of his doctrine, at any rate, fairly challenge. It was chiefly 
because he seemed to make freedom * an unrealised and un- 
realisable state, that his moral doctrine was found un- 
satisfactory by Hegel. Hegel holds that freedom, as the 
condition in which the will is determined by an object 
adequate to itself, or by an object which itself as reason 
constitutes, is realised in the state. He thinks of the state 
in a way not familiar to Englishmen, a way not unlike that 
in which Greek philosophers thought of the rrroXis, as a society 
governed by laws and institutions and established customs 
which secure the common good of the members of the society 
—enable them to make the best of themselves — and are re- 
cognised as doing so. Such a state is ' objective freedom ' ; 
freedom is realised in it because in it the reason, the self- 
determining principle operating in man as his will, has fonnd 
a perfect expression for itself (as an artist maybe considered 
to express himself in a perfect work of art) ; and the man 
who is determined by the objects which the well-ordered 
state presents to him is determined by that which is the 
perfect expression of his reason, and is thus free. 

5. There is, no doubt, truth in this view. I have already 
tried to show 2 how the self-distinguishing and self-seeking 
consciousness of man, acting in and upon those human wants 
and ties and affections which in their proper human character 
have as little reality apart from it as it apart from them, 
gives rise to a system of social relations, with laws, 
customs, and institutions corresponding; and how in this 
system the individual's consciousness of the absolutely desir- 
able, of something that should be, of an ideal to be realised 
in his life, finds a content or object which has been 
constituted or brought into being by that consciousness 
itself as working through generations of men ; how interests 
are thus supplied to the man of a more concrete kind than 

1 In the sense of autonomy of rational determination Kant would have recog- 

will,' or determination by an object nised as characteristic of every human 

which reason constitutes, as distinct act, properly so called, 

from determination by an object which z [In a previous course of lectures, 

the man makes his own; this latter See Prolegomena to Ethic*, III. iii.] 


the interest in fulfilment of a universally binding law 
because universally binding, but which yet are the product of 
reason, and in satisfying which he is conscious of attaining a 
true good, a good contributory to the perfection of himself and 
his kind. There is thus something in all forms of society that 
tends to the freedom l at least of some favoured individuals, 
because it tends to actualise in them the possibility of that 
determination by objects conceived as desirable in distinction 
from objects momentarily desired, which is determination by 
reason. 2 To put it otherwise, the effect of his social relations 
on a man thus favoured is that, whereas in all willing the 
individual seeks to satisfy- himself, this man seeks to satisfy 
himself, not as one who feels this or that desire, but as one 
who conceives, whose nature demands, a permanent good. 
So far as it is thus in respect of his rational nature that he 
makes himself an object to himself, his will is autonomous. 
This was the good which the ideal ttoXls, as conceived by 
the Greek philosophers, secured for the true 7ro\lrr)9, the 
man who, entering into the idea of the ttoXis, was equally 
qualified apyziv /cal ap^saOai. No doubt in the actual Greek 
ttoXls there was some tendency in this direction, some 
tendency to rationalise and moralise the citizen. With- 
out the real tendency the ideal possibility would not 
have suggested itself. And in more primitive forms of 
society, so far as they were based on family or tribal 
relations, we can see that the same tendency must have been 
at work, just as in modern life the consciousness of his 
position as member or head of a family, wherever it exists, 
necessarily does something to moralise a man. In modern 
Christendom, with the extension of citizenship, the security 
of family life to all men (so far as law and police can secure 
it), the establishment in various forms of Christian fellowship 
of which the moralising functions grow as those of the 
magistrate diminish, the number of individuals whom society 
awakens to interests in objects contributory to human per^ 
fection tends to increase. So far the modern state, in that 
full sense in which Hegel uses the term (as including all the 
agencies for common good of a law-abiding people), does 
contribute to the realisation of freedom, if by freedom we 
understand the autonomy of the will or its determination by 

1 In the sense of ' autonomy of will.' 

2 [This last clause is queried in the MS.] 


rational objects, objects which help to satisfy the demand 
of reason, the effort after self-perfection. 

6. On the other hand, it would seem that we cannot 
significantly speak of freedom except with reference to 
individual persons ; that only in them can freedom be 
realised ; that therefore the realisation of freedom in the 
state can only mean the attainment of freedom by indi- 
viduals through influences which the state (in the wide 
sense spoken of) supplies, — * freedom ' here, as before, 
meaning not the mere self-determination which renders us 
responsible, but determination by reason, ' autonomy of the 
will ' ; and that under the best conditions of any society 
that has ever been such realisation of freedom is most 
imperfect. To an Athenian slave, who might be used to 
gratify a master's lust, it would have been a mockery to 
speak of the state as a realisation of freedom ; and perhaps 
it would not be much less so to speak of it as such to an 
untaught and under-fed denizen of a London yard with 
gin-shops on the right hand and on the left. What Hegel 
says of the state in this respect seems as hard to square 
with facts as what St. Paul says of the Christian whom the 
manifestation of Christ has transferred from bondage into 
' the glorious liberty of the sons of God.' In both cases the 
difference between the ideal and the actual seems to be 
ignored, and tendencies seem to be spoken of as if they 
were accomplished facts. It is noticeable that by uncritical 
readers of St. Paul the account of himself as under the law 
(in Romans vii.), with the ' law of sin in his members warring 
against the law of his reason,' is taken as applicable to the 
regenerate Christian, though evidently- St. Paul meant it as 
a description of the state from which the Gospel, the 
* manifestation of the Son of God in the likeness of sinful 
flesh,' set him free. They are driven to this interpretation 
because, though they can understand St. Paul's account of 
his deliverance as an account of a deliverance achieved for 
them but not in them, or as an assurance of what is to be, 
they cannot adjust it to the actual experience of the 
Christian life. In the same way Hegel's account of freedom 
as realised in the state does not seem to correspond to the 
facts of society as it is, or even as, under the unalterable 
conditions of human nature, it ever could be; though 
undoubtedly there is a work of moral liberation, which 


society, through its various agencies, is constantly carrying 
on for the individual. 

7. Meanwhile it must be borne in mind that in all these 
different views as to the manner and degree in which 
freedom is to be attained, ' freedom ' does not mean that 
the man or will is undetermined, nor yet does it mean mere 
self-determination, which (unless denied altogether, as by 
those who take the strictly naturalistic view of human 
action) must be ascribed equally to the man. whose will is 
heteronomous or vicious, and to him whose will is auto- 
nomous ; equally to the man who recognises the authority 
of law in what St. Paul would count the condition of a 
bondman, and to him who fulfils the righteousness of the 
law in the spirit of adoption. It means a particular kind of 
self-determination ; the state of the man who lives indeed 
for himself, but for the fulfilment of himself as a ' giver of 
law universal ' (Kant) ; who lives for himself, but only 
according to the true idea of himself, according to the law 
of his being, ' according to nature ' (the Stoics) ; who is so 
taken up into God, to whom God so gives the spirit, that 
there is no constraint in his obedience to the divine will 
(St. Paul) ; whose interests, as a loyal citizen, are those of a\ 
well-ordered state in which practical reason expresses 1 
itself (Hegel). Now none of these modes of self-deter-/ 
mination is at all implied in ' freedom ' according to the| 
primary meaning of the term, as expressing that relation 
between one man and others in which he is secured from 
compulsion. All that is so implied is that a man should 
have power to do what he wills or prefers. No reference is 
made to the nature of the will or preference, of the object 
willed or preferred; whereas according to the usage oft 
* freedom ' in the doctrines we have just been considering, itl 
is not constituted by the mere fact of acting upon preference^ 
but depends wholly on the nature of the preference, upon) 
the kind of object willed or preferred. 

8. If it were ever reasonable to wish that the usage of 
words had been other than it has been (any more than that 
the processes of nature were other than they are), one might 
be inclined to wish that the term 'freedom' had been con- 
fined to the juristic sense of the power to c do what one wills ' : 
for the extension of its meaning seems to have caused much 
controversy and confusion. But, after all, this extension 


does but represent various stages of reflection upon the 
self-distinguishing, self-seeking, self-asserting principle, of 
which the establishment of freedom, as a relation between 
man and man, is the expression. The reflecting man is not 
content with the first announcement which analysis makes 
as to the inward condition of the free man, viz. that he can 
do what he likes, that he has the power of acting according 
to his will or preference. In virtue of the same principle 
which has led him to assert himself against others, and thus 
to cause there to be such a thing as (outward) freedom, he 
distinguishes himself from his preference, and asks how he is 
related' to it, whether he determines it or how it is deter- 
mined. Is he free to will, as he is free to act ; or, as the 
act is determined by the preference, is the preference deter- 
mined by something else? Thus Locke (Essay, II. 21) begins 
with deciding that freedom means power to do or forbear 
from doing any particular act upon preference, and that, 
since the will is merely the power of preference, the question 
whether the will is free is an unmeaning one (equivalent to 
the question whether one power has another power) ; that 
thus the only proper question is whether a man (not his will) 
is free, which must be answered affirmatively so far as he 
has the power to do or forbear, as above. But he recognises 
the propriety of the question whether a man is free to will 
as well as to act. He cannot refuse to carry back the 
analysis of what is involved in a man's action beyond the 
preference of one possible action to another, and to inquire 
what is implied in the preference. It is when this latter 
question is raised, that language which is appropriate enough 
in a definition of outward or juristic freedom becomes mis- 
leading. It having been decided that the man civilly free 
has power over his actions, to do or forbear according to 
preference, it is asked whether he has also power to prefer. 

9. But while it is proper to ask wdiether in any particular 
case a man has power over his actions, because his nerves and 
limbs and muscles may be acted upon by^ another person or 
a force which is not he or his, there is no appropriateness in 
asking the question in regard to a preference or will, because 
this cannot be so acted on. If so acted on, it would not be 
a will or preference. There is no such thing as a will which 
a man is not conscious of as belonging to himself, no such 
thing as an act of will which he is not conscious of as 


issuing from himself. To ask whether he has power over it, 
or whether some other power than he determines it, is like 
asking whether he is other than himself. Thus the question 
whether a man, having power to act according to his will, 
or being free to act, has also power over his will, or is free 
to will, has just the same impropriety that Locke points out 
in the question whether the will is free. The latter question, 
on the supposition that there is power to enact the will, — a 
supposition which is necessarily made by those who raise the 
ulterior question whether there is power over the will, — is 
equivalent, as Locke sees, to a question whether freedom is 
free. For a will which there is power of enacting consti- 
tutes freedom, and therefore to ask whether it is free is like 
asking (to use Locke's instance) whether riches are rich 
(' rich' being a denomination from the possession of riches, 
just as 'free ' is a denomination from the possession of free- 
dom, in the sense of a will which there is power to enact). 
But if there is this impropriety in the question whether the 
will is free, there is an equal one in the question which 
Locke entertains, viz. whether man is free to will, or has 
power over his will. It amounts to asking whether a cer- 
tain power is also a power over itself : or, more precisely, 
whether a man possessing a certain power — that which we 
call freedom — has also the same power over that power. 

10. It may be said perhaps that we are here pressing 
words too closely ; that it is of course understood, when it is 
asked whether a man has power over his will, that ' power ' 
is used in a different sense from 'that which it bears when it 
is asked whether he has power to enact his will : that ' free- 
dom,' in like manner, is understood to express a different 
kind of power or relation when we ask whether a man is 
free to will, and when we ask whether he is free to act. But 
granting that all this has been understood, the misleading 
effects of the question in the form under consideration (' Is a 
man free to will as well as to act ? ' * Has he power over his 
will ? ') remain written in the history of the ' free-will con- 
troversy.' It has mainly to answer for two wrong ways of 
thinking on the subject; (a) for the way of thinking of the 
determining motive of an act of will, the object willed, as 
something apart from the will or the man willing, so that in 
being determined by it the man is supposed not to be self- 
determined, but to be determined as one natural event by 


another, or at best as a natural organism by the forces 
acting on it : (b), for the view that the only way of escaping 
this conclusion is to regard the will as independent of 
motives, as a power of deciding between motives without 
any motive to determine the decision, which must mean 
without reference to any object willed. A man, having (in 
virtue of his power of self-distinction and self-objectification) 
presented his will to himself as something to be thought 
about, and being asked whether he has power over it, 
whether he is free in regard to it as he is free against other 
persons and free to use his limbs and, through them, 
material things, this way or that, must very soon decide that 
he is not. His will is himself. His character necessarily 
shows itself in his will. We have already, in a previous 
lecture, 1 noticed the practical fallacy involved in a man's 
saying that he cannot help being what he is, as if he were 
controlled by external power ; but he being what he is, and 
the circumstances being what they are at any particular con- 
juncture, the determination of the will is already given, just 
as an effect is given in the sum of its conditions. The deter- 
mination of the will might be different, but only through the 
man's being different. But to ask whether a man has power 
o\er determinations of his will, or is free to will as he is to 
act, as the question is commonly understood and as Locke 
understood it, is to ask whether, the man being what at any 
time he is, it is still uncertain (1) whether he will choose or 
forbear choosing between certain possible courses of action, 
and (2) supposing him to choose one or other of them, which 
he will choose. 

11. Now we must admit that there is really no such 
uncertainty. The appearance of it is due to our ignorance 
of the man and the circumstances. If, however, because this 
is so, we answer the question whether a man has power over 
his will, or is free to will, in the negative, 2 we at once 
suggest the conclusion that something else has power over 
it, viz. the strongest motive. We ignore the truth that in 
being determined by a strongest motive, in the only sense 
in which he is really so determined, the man (as previously 

> [Prolegomena to Ethics, §§ 107, ff-] since a man's will is himself, and 

2 Instead of saying (as we should) 'freedom' and 'power' express rela- 

that it is one of t those inappropriate tions between a man and something 

questions to which there is no answer ; other than himself. 


explained) l is determined by himself, by an object of his 
own making, and we come to think of the will as determined 
like any natural phenomenon by causes external to it. All 
this is the consequence of asking questions about the 
relation between a man and his will in terms only appro- 
priate to the relation between the man and other men, or 
to that between the man and his bodily members or the 
materials on which he acts through them. 

12. On the other side the consciousness of self-determina- 
tion resists this conclusion ; but so long as we start from the 
question whether a man has power over his will, or is free 
to will as well as to act, it seems as if the objectionable 
conclusion could only be avoided by answering this question 
in the affirmative. But to say that a man has power over 
determinations of his will is naturally taken to mean that 
he can change his will while he himself remains the same ; 
that given his character, motives, and circumstances as these 
at any time are, there is still something else required for 
the determination of his will ; that behind and beyond the 
will as determined by some motive there is a will, itself un- 
determined by any motive, that determines what the deter- 
mining motive shall be, — that - has power over ' his preference 
or choice, as this has over the motion of his bodily members. 
But an unmotived will is a will without an object, which is 
nothing. The power or possibilitj^, beyond any actual deter- 
mination of the will, of determining what that determination 
shall be is a mere negation of the actual determination. It 
is that determination as it becomes after an abstraction of 
the motive or object willed, which in fact leaves nothing at 
all. If those moral interests, which are undoubtedly in- 
volved in the recognition of the distinction between man and 
any natural phenomenon, are to be made dependent on belief 
in such a power or abstract possibility, the case is hopeless. 

13. The right way out of the difficulty lies in the dis- 
cernment that the question whether a man is free to will, or 
has power over the determinations of his will, is a question to 
which there is no answer, because it is asked in inappropriate 
terms ; in terms that imply some agency beyond the will 
which determines what the will shall be (as the will itself is 
an agency beyond the motions of the muscles which deter- 
mines what those motions shall be), and that as to this 

1 [See Prolegomena to Ethics, § 105.] 


agency it may be asked whether it does or does not lie in the 
man himself. In trnth there is no such agency beyond the 
will and determining how the will shall be determined ; not 
in the man, for the will is the self-conscious man ; not else- 
where than in the man, not outside him, for the self-conscious 
man has no outside. He is not a body in space with other 
bodies elsewhere in space acting upon it and determining 
its motions. The self-conscious man is determined by 
objects, which in order to be objects must already be in con- 
sciousness, and in order to be his objects, the objects which 
determine him, must already have been made his own. To 
say that they have power over him or his will, and that he 
or his will has power over them, is equally misleading. 
Such language is only applicable to the relation between an 
agent and patient, when the agent and the patient (or at any 
rate the agent) can exist separately. But self-consciousness 
and its object, will and its object, form a single individual 
unity. Without the constitutive action of man or his will 
the objects do not exist ; apart from determination by some 
object neither he nor his will would be more than an unreal 

14. If, however, the question is persisted in, ' Has a man 
power over the determinations of his will ? ' "we must 
answer both ' yes ' and 'no.' ' No,' in the sense that he is 
not other than his will, with ability to direct it as the will 
directs the muscles. ' Yes,' in the sense that nothing ex- 
ternal to him or his will or self-consciousness has power over 
them. ' No,' again, in the sense that, given the man and 
his object as he and it at anytime are, there is no possibility 
of the will being determined except in one way, for the will 
is already determined, being nothing else than the man as 
directed to some object. ' Yes,' in the sense that; the deter- 
mining object is determined "by the man or will just as much 
as the man or will by the object. The fact that the state of 
the man, on which the nature of his object at any time 
depends, is a result of previous states, does not affect the 
validity of this last assertion, since (as we have seen ') all 
these states are states of a self-consciousness from which all 
alien determination, all determination except through the 
medium of self-consciousness, is excluded. 

15. In the above we have not supposed any account to be 

1 [Prolegomena to Ethic*, § 102.] 


taken of the character of the objects willed in the application 
to the will itself of the qnestion ' free or not free/ which is 
properly applied only to an action (motion of the bodily 
members) or to a relation between one man and other men. 
Those who unwisely consent to entertain the question whether 
a man is free to will or has power over determinations of his 
will, and answer it affirmatively or negatively, consider their 
answer, whether ( yes ' or ' no,' to be equally applicable what- 
ever the nature of the objects willed. If they decide that a 
man is ' free to will,' they mean that he is so in all cases of 
willing, whether the object willed be a satisfaction of animal 
appetite or an act of heroic self-sacrifice ; and conversely, if 
they decide that he is not free to will, they mean that he is not 
so even in cases when the action is done upon cool calculation or 
upon a principle of duty, as much as when it is done on im- 
pulse or in passion. Throughout the controversy as to free 
will that has been carried on among English psychologists 
this is the way in which the question has been commonly dealt 
with. The freedom, claimed or denied for the will, has been 
claimed or denied for it irrespectively of those objects willed, 
on the nature of which the goodness or badness of the will 

16. On the other hand, with the Stoics, St. Paul, Kant, 
and Hegel, as we have seen, the attainment of freedom (at 
any rate of the reality of freedom, as distinct from some 
mere possibility of it which constitutes the distinctive human 
nature) depends on the character of the objects willed. In 
all these ways of thinking, however variously the proper object 
of will is conceived, it is only as directed to this object, and 
thus (in Hegelian language) Corresponding to its idea, that 
the will is supposed to be free. The good will is free, not 
the bad will. Such a view of course implies some element 
of identity between good will and bad will, between will as 
not yet corresponding to its idea and will as so correspond- 
ing. St. Paul indeed, not being a systematic thinker and 
being absorbed in the idea of divine grace, is apt to speak as 
if there were nothing in common between the carnal or natural 
man (the will as in bondage to the flesh) and the spiritual 
man (the will as set free) ; just as Plato commonly ignores 
the unity of principle in all a man's actions, and repre- 
sents virtuous actions as coming from the God in man, 
vicious actions from the beast. Kant and Hegel, however,— 



though they do not consider the will as it is in every man, 
good and bad, to be free ; though Kant in his later ethical 
writings, and Hegel (I think) always, confine the term 
4 Wille ' to the will as having attained freedom or come to 
correspond to its idea, and apply the term ' Willkiir ' to that 
self- determining principle of action which belongs to every 
man and is in their view the mere possibility, not actuality, of 
freedom, — yet quite recognise what has been above insisted on 
as the common characteristic of all willing, the fact that it is 
not a determination from without, like the determination of 
any natural event or agent, but the realisation of an object 
which the agent presents to himself or makes his own, the 
determination by an object of a subject which itself consciously 
determines that object ; and they see that it is only for a sub- 
ject free in this sense ('an sich' but not ' fur sich,' hwa/xsu 
but not svspyeia) that the reality of freedom can exist. 

17. Now the propriety or impropriety of the use of 

'freedom ' to express the state of the will, not as directed to any 

and every object, but only to those to which, according to the 

law of nature or the will of God or its 'idea,' it should be 

directed, is a matter of secondary importance. This usage 

of the term is, at any rate, no more a departure from the 

primary or juristic sense than is its application to the will as 

distinct from action in any sense whatever. And certainly the 

jun sophisticated man, as soon as the usage of ' freedom ' 

to express exemption from control by other men and ability 

to do as he likes is departed from, can much more readily 

assimilate the notion of states of the inner man described 

as bondage to evil passions, to terrors of the law, or on 

the other hand as freedom from sin and law, freedom in 

the consciousness of union with God, or of harmony with the 

trne law of one's being, freedom of true loyalty, freedom 

in devotion to self-imposed duties, than he can assimilate 

the notion of freedom as freedom to will anything and 

everything, or as exemption from determination by motives, 

or the constitution by himself of the motives which determine 

his will. And there is so far less to justify the extension 

of the usage of the term in these latter ways than in the 

1 former. It would seem indeed that there is a real community 

of meaning between ' freedom ' as expressing the condition of 

a citizen of a civilised state, and ' freedom ' as expressing 

the condition of a man who is inwardly ' master of himself.' 


That is to say, the practical conception by a man (' practical ' 
in the sense of having a tendency to realise itself) of a self- 
satisfaction to be attained in his becoming what he should 
be, what he has it in him to be, in fulfilment of the law of 
his being, — or, to vary the words but not the meaning, in 
attainment of the righteousness of God, or in perfect obedi- 
ence to self-imposed law, — this practical conception is the 
outcome of the same self-seeking principle which appears in 
a man's assertion of himself against other men and against 
nature p against other men,' as claiming their recognition of 
him as being what they are ; c against nature,' as able to use it). 
This assertion of himself is the demand for freedom, freedom 
in the primary or juristic sense of power to act according to 
choice or preference. So far as such, freedom is established 
for any man, this assertion of himself is made good; and 
such freedom is precious to him because it is an achieve- 
ment of the self-seeking principle. It is a first satisfaction 
of its claims, which is the condition of all other satisfaction 
of them. The consciousness of it is the first form of self- 
enjoyment, of the joy of the self-conscious spirit in itself as 
in the one object of absolute value. 

18. This form of self-enjoyment, however, is one which 
consists essentially in the feeling by the subject of a possi- 
bility rather than a reality, of what it has it in itself to 
become, not of what it actually is. To a captive on first 
winning his liberty, as to a child in the early experience of 
power over his limbs and through them over material things, 
this feeling of a boundless possibility of becoming may give 
real joy ; but gradually the sense of what it is not, of the 
very little that it amounts to, must predominate over the 
sense of actual good as attained in it. Thus to the grown 
man, bred to civil liberty in a society which has learnt to 
make nature its instrument, there is no self-enjoyment in 
the mere consciousness of freedom as exemption from external 
control, no sense of an object in which he can satisfy himself 
having been obtained. 

Still, just as the demand for and attainment of freedom 
from external control is the expression of that same self- 
seeking principle from which the quest for such an object 
proceeds, so 'freedom' is the natural term by which the 
man describes such an object to himself, — describes to him- 
self the state in which he shall have realised his ideal of 

T 2 


himself, shall be at one with the law which he recognises as 
that which he ought to obey, shall have become all that he 
has it in him to be, and so fulfil the law of his being or ' live 
according to nature.' Just as the consciousness of an 
unattainable ideal, of a law recognised as having authority 
but with which one's will conflicts, of wants and impulses 
which interfere with the fulfilment of one's possibilities, is a 
consciousness of impeded energy, a consciousness of oneself 
as for ever thwarted and held back, so the forecast of 
deliverance from these conditions is as naturally said to be 
a forecast of ' freedom ' as of c peace ' or ' blessedness.' Nor 
is it merely to a select few, and as an expression for a 
deliverance really (as it would seem) unattainable under the 
conditions of any life that we know, but regarded by saints 
as secured for them in another world, and by philosophers 
as the completion of a process which is eternally complete 
in God, that ' freedom ' commends itself. To any popular 
audience interested in any work of self-improvement (e.g. 
to a temperance- meeting seeking to break the bondage to 
liquor), it is as an effort to attain freedom that such work 
can be most effectively presented. It is easy to tell such 
people that the term is being misapplied ; that they are 
quite ( free ' as it is, because every one can do as he likes 
so long as he does not prevent another from doing so ; 
that in any sense in which there is such a thing as ' free 
will,' to get drunk is as much an act of free will as any- 
thing else. Still the feeling of oppression, which always 
goes along with the consciousness of unfulfilled possibili- 
ties, will always give meaning to the representation of the 
effort after any kind of self-improvement as a demand for 
6 freedom.' 

19. The variation in the meaning of ' freedom ' having 
been thus recognised and accounted for, we come back to the 
more essential question as to the truth of the view which 
underlies all theories implying thatjreedom is in some sense 
the goal of moral endeavour ; the view, namely, that there 
is some will in a man with which many or most of his volun- 
tary actions do not accord, a higher self that is not satisfied 
by ~the objects which yet he deliberately pursues. Some 
such notion is common to those different theories about free- 
dom which in the rough we have ascribed severally to the 
Stoics, St. Paul, Kant, and Hegel. It is the same notion 


which was previously l put in the form, ' that a man is sub- 
ject to a law of his being, in virtue of which he at once seeks 
self-satisfaction, and is prevented from finding it in the 
objects which he actually desires, and in which he ordinarily 
seeks it.' 'What can this mean?' it maybe asked. 'Of 
course we know that there are weak people who never suc- 
ceed in getting what they want, either in the sense that they 
have not a,bility answering to their will, or that they are 
always wishing for something which yet they do not will. 
But it would not be very appropriate to apply the above 
formula to such people, for the man's will to attain certain 
objects cannot be ascribed to the same law of his being as 
the lack of ability to attain them, nor his wish for certain 
objects to the same law of his being as those stronger desires 
which determine his will in a contrary direction. At any 
rate, if the proposition is remotely applicable to the man 
who is at once selfish and unsuccessful, how can it be true 
in any sense either of the man who is at once selfish and 
succeeds, who gets what he wants (as is unquestionably the 
case with many people who live for what a priori moralists 
count unworthy objects), or of the man who 'never thinks 
about himself at all'? So far as the proposition means any- 
thing, it would seem to represent Kant's notion, long ago 
found unthinkable and impossible, the notion of there being 
two wills or selves in a man, the ' pure' will or ego and the 
' empirical ' will or ego, the pure will being independent of a 
man's actual desires and directed to the fulfilment of a uni- 
versal law of which it is itself the giver, the empirical will 
being determined by the strongest desire and directed to this 
or that pleasure. In this proposition the * objects which the 
man actually desires and in which he ordinarily seeks satis- 
faction ' are presumably objects of what Kant called the 
'empirical will,' while the 'law of his being' corresponds to 
Kant's ' pure ego.' But just as Kant must be supposed to 
have believed in some identity between the pure and em- 
pirical will, as implied in the one term 'will,' though he 
does not explain in what this identity consists, so the pro- 
position before us apparently ascribes man's quest for self- 
satisfaction as directed to certain objects, to the same law of 
his being which prevents it from finding it there. Is not 
this nonsense ? ' 

1 [Above, section 1.] 


20. To such questions we answer as follows. The pro- 
position before us, like all the theories of moral freedom 
which we have noticed, undoubtedly implies that the will 
of every man is a form of one consciously self-realising 
principle, which at the same time is not truly or fully ex- 
pressed in any man's will. As a form of this self-realising 
principle it may be called, if we like, a ' pure ego ' or ' the 
pure ego ' of the particular person ; as directed to this or that 
object in such a way that it does not truly express the self- 
realising principle of which it is a form, it may be called the 
' empirical ego ' of that person. But if we use such language, 
it must be borne in mind that the pure and empirical egos 
are still not two egos but one ego ; the pure ego being the 
self-realising principle considered with reference either to its 
idea, its possibility, what it has in itself to become, the law 
of its being, or to some ultim ate actualisation of this possibility ; 
the empirical ego being the same principle as it appears in 
this or that state of character, which results from its action, 
but does not represent that which it has in itself to become, 
does not correspond to its idea or the law of its being. By 
a consciously self-realising principle is meant a principle 
that is determined to action by the conception of its own 
perfection, or by the idea of giving reality to possibilities 
which are involved in it and of which it is conscious as so 
involved ; or, more precisely, a principle which at each stage 
of its existence is conscious of a more perfect form of exist- 
ence as possible for itself, and is moved to action by that 
consciousness. We must now explain a little more fully how 
we understand the relation of the principle in question to 
what we call our wills and our reason,-^-the will and reason 
of this man and that, — and how we suppose its action to con- 
stitute the progress of morality. 

21. By i practical reason ' we mean a consciousness of a 
possibility of perfection to be realised in and by the subject 
of the consciousness. By ' will ' we mean the effort of a self- 
conscious subject to satisfy itself. In God, so far as we can 
ascribe reason and will to Him, we must suppose them to 
be absolutely united. In Him there can be no distinction 
between possibility and realisation, between the idea of 
perfection and the activity determined by it. But in men 
the self-realising principle, which is the manifestation of 
God in the world of becoming, in the form which it takes 


as will at best only tends to reconciliation with, itself in the 
form which it takes as reason. Self-satisfaction, the pursuit 
of which is will, is sought elsewhere than in the realisation 
of that consciousness of possible perfection, which is reason. 
In this sense the object of will does not coincide with the 
object of reason. On the other hand, just because it is self- 
satisfaction that is sought in all willing, and because by a 
self-conscious and self-realising subject it is only in the 
attainment of its own perfection that such satisfaction can 
be found, the object of will is intrinsically or potentially, 
and tends to become actually, the same as that of reason. It 
is this that we express by saying that man is subject to a 
law of his being which prevents him from finding satisfaction 
in the objects in which under the pressure of his desires it is 
his natural impulse to seek it. This ' natural impulse ' (not 
strictly 'natural') is itself the result of the operation of the 
self -realising principle upon what would otherwise be an 
animal system, and is modified, no doubt, with endless com- 
plexity in the case of any individual by the result of such 
operation through the ages of human history. But though 
the natural impulses of the will are thus the work of the self- 
realising principle in us, it is not in their gratification that 
this principle can find the satisfaction which is only to be 
found in the consciousness of becoming perfect, of realising 
what it has it in itself to be. In order to any approach to 
this satisfaction of itself the self-realising principle must 
carry its work farther. It must overcome the ' natural 
impulses,' not in the sense of either extinguishing them or 
denying them an object, but in the sense of fusing them 
with those higher interests, which have human perfection 
in some of its forms for their object. Some approach to 
this fusion we may notice in all good men ; not merely in 
those in whom all natural passions, love, anger, pride, am- 
bition, are enlisted in the service of some great public cause, 
but in those with whom such passions are all governed 
by some such commonplace idea as that of educating a 

22. So far as this state is reached, the man may be said 
to be reconciled to ' the law of his being ' which (as was 
said above) prevents him from finding satisfaction in the 
objects in which he ordinarily seeks it, or anywhere but in 
the realisation in himself of an idea of perfection. Since the 


law is, in fact, the action of that self-realising subject which 
is his self, and which exists in God as eternally self-realised, 
he may be said in this reconciliation to be at peace at once 
with himself and with God. 

Again, he is ' free,' (1) in the sense that he is the author 
of the law which he obeys (for this law is the expression of 
that which is his self), and that he obeys it because 
conscious of himself as its author; in other words, obeys it 
from that impulse after self-perfection which is the source 
of the law or rather constitutes it. He is ' free ' (2) in the 
sense that he not merely ' delights in the law after the 
inward man ' (to use St. Paul's phrase), while his natural 
impulses are at once thwarted by it and thwart him in his 
effort to conform to it, but that these very impulses have 
been drawn into its service, so that he is in bondage neither 
to it nor to the flesh. 

From the same point of view we may say that his will is 
'autonomous,' conforms to the law which the will itself consti- 
tutes, because the law (which prevents him from finding satis- 
faction anywhere but in the realisation in himself of an idea 
of perfection) represents the action in him of that self- 
realising principle of which his will is itself a form. There 
is an appearance of equivocation, however, in this way of 
speaking, because the * will ' which is liable not to be autono- 
mous, and which we suppose gradually to approach autonomy 
in the sense of conforming to the law above described, is 
not this self-realising principle in the form in which this 
principle involves or gives the law. On the contrary, it 
is the self-realising principle as constituting that effort 
after self-satisfaction in each of us which is liable to be and 
commonly is directed to objects which are not contributory 
to the realisation of the idea of perfection, — objects which 
the self-realising principle accordingly, in the fulfilment of 
its work, has to set aside. The equivocation is pointed out by 
saying, that the good will is * autonomous ' in the sense of 
conforming to a law which the will itself, as reason, constitutes ; 
which is, in fact, a condensed way of saying, that the good 
will is the will of which the object coincides with that of 
practical reason ; that will has its source in the same self- 
realising principle which yields that consciousness of a 
possible self-perfection which we call reason,and that it can 
only correspond to its idea, or become what it has the possi- 


bility of becoming, in being directed to the realisation of that 

23. According to the view here taken, then, reason and 
will, even as they exist in men, are one in the sense that they 
are alike expressions of one self-realising principle. In God, 
or rather in the ideal human person as he really exists in 
God, they are actually one ; i.e. self-satisfaction is for ever 
sought and found in the realisation of a completely articulated 
or thoroughly filled idea of the perfection of the human person. 
In the historical man — in the men that have been and are 
coming to be — they tend to unite. In the experience of 
mankind, and again in the experience of the individual as 
determined by the experience of mankind, both the idea of 
a possible perfection of man, the idea of which reason is the 
faculty, and the impulse after self-satisfaction which belongs 
to the will, undergo modifications which render their recon- 
ciliation in the individual (and it is only in individuals that 
they can be reconciled, because it is only in them that they 
exist) more attainable. These modifications may be stated 
summarily as (1) an increasing concreteness in the idea of 
human perfection ; its gradual development from the vague 
inarticulate feeling that there is such a thing into a concep- 
tion of a complex organisation of life, with laws and institu- 
tions, with relationships, courtesies, and charities, with arts 
and graces through which the perfection is to be attained ; 
and (2) a corresponding discipline, through inheritance and 
education, of those impulses which ina,y be called ' natural ' 
in the sense of being independent of any conscious direction 
to the fulfilment of an idea of perfection. Such discipline 
does not amount to the reconciliation of will and reason ; it 
is not even, properly speaking, the beginning of it ; for the 
reconciliation only begins with the direction of the impulse 
after self-satisfaction to the realisation of an idea of what 
should be, as such (because it should be) ; and no discipline 
through inheritance or education, just because it is only 
impulses that are natural (in the sense defined) which it can 
affect, can bring about this direction, which, in theological 
language, must be not of nature, but of grace. On the con- 
trary, the most refined impulses may be selfishly indulged ; 
i.e. their gratification may be made an object in place of that 
object which consists in the realisation of the idea of per- 
fection. But unless a discipline and refinement of the natural 


impulses, through the operation of social institutions and arts, 
went on pari passu with the expression of the idea of perfection 
in such institutions and arts, the direction of the impulses of 
the individual by this idea, when in some form or other it 
has been consciously awakened in him, would be practically 
impossible. The moral progress of mankind has no reality 
except as resulting in the formation of more perfect indi- 
vidual characters ; but on the other hand every progress 
towards perfection on the part of the individual character 
presupposes some embodiment or expression of itself by the 
self-realising principle in what may be called (to speak most 
generally) the organisation of life. It is in turn, however, 
only through the action of individuals that this organisation 
of life is achieved. 

24. Thus the process of reconciliation between will and 
reason, — the process through which each alike comes actually 
to be or to do what it is and does in possibility, or according 
to its idea, or according to the law of its being, — so far as 
it comes within our experience may be described as follows. 
A certain action of the self-realising principle, of which 
individuals susceptible in various forms to the desire to 
better themselves have been the media, has resulted in con- 
ventional morality ; in a system of recognised rules (whether 
in the shape of law or custom) as to what the good of society 
requires, which no people seem to be wholly without. The 
moral progress of the individual, born and bred under such a 
system of conventional morality, consists (1) in the adjust- 
ment of the self-seeking principle in him to the requirements 
of conventional morality, so that the modes in which he 
seeks self-satisfaction are regulated by . the sense of what is 
expected of him. This adjustment (which it is the business 
of education to effect) is so far a determination of the will 
as in the individual by objects which the universal or 
national human will, of which the will of the individual is a 
partial expression, has brought into existence, and is thus 
a determination of the will by itself. It consists (2) in a 
process of reflection, by which this feeling in the individual 
of what is expected of him becomes a conception (under 
whatever name) of something that universally should be, of 
something absolutely desirable, of a single end or object of 
life. The content of this conception may be no more than 
what was already involved in the individual's feeling of what 


is expected of him ; that is to say, if called upon to state in 
detail what it is that has to be done for the attainment of 
the absolute moral end or in obedience to the law of what 
universally should be, he might only be able to specify con- 
duct which, apart from any such explicit conception, he felt 
was expected of him. For all that there is a great difference 
between feeling that a certain line of conduct is expected of 
me and conceiving it as a form of a universal duty. So long 
as the requirements of established morality are felt in the 
former way, they present themselves to the man as imposed 
from without. Hence, though they are an expression of 
practical reason, as operating in previous generations of 
men, yet, unless the individual conceives them as relative to 
an absolute end common to him with all men, they become 
antagonistic to the practical reason which operates in him, 
and which in him is the source at once of the demand for 
self-satisfaction and of the effort to find himself in, to carry 
his own unity into, all things presented to him. Unless the 
actions required of him by 'the divine law, the civil law, and 
the law of opinion or reputation ' (to use Locke's classifica- 
tion) tend to realise his own idea of what should be or is good 
on the whole, they do not form an object which, as contem- 
plated, he can harmonise with the other objects which he 
seeks to understand, nor, as a practical object, do they form 
one in the attainment of which he can satisfy himself. Hence 
before the completion of the process through which the in- 
dividual comes to conceive the performance of the actions 
expected of him under the general form of a duty which in 
the freedom of his own reason he recognises as binding, 
there is apt to occur a revolt against conventional morality. 
The issue of this may either be an apparent suspension of the 
moral growth of the individual, or a clearer apprehension of 
the spirit underlying the letter of the obligations laid on him 
by society, which makes his rational recognition of duty, 
when arrived at, a much more valuable influence in promot- 
ing the moral growth of society. 

25. Process (2), which may be called a reconciliation of 
reason with itself, because it is the appropriation by reason 
as a personal principle in the individual of the work which 
reason, acting through the media of other persons, has already 
achieved in the establishment of conventional morality, is the 
condition of the third stage in which the moral progress of 


the individual consists ; viz. the growth of a personal interest 
in the realisation of an idea of what should be, in doing what 
is believed to contribute to the absolutely desirable, or to 
human perfection, because it is believed to do so. Just so 
far as this interest is formed, the reconciliation of the two 
modes in which the practical reason operates in the individual 
is effected. The demand for self-satisfaction (practical reason 
as the will of the individual) is directed to the realisation of 
an ideal object, the conceived ' should be,' which practical 
reason as- our reason constitutes. The c autonomy of the 
will' is thus attained in a higher sense than it is in the 
' adjustment ■ described under (1), because the objects to 
which it is directed are not merely determined by customs and 
institutions which are due to the operation of practical reason 
in previous ages, but are embodiments or expressions of the 
conception of what absolutely should be as formed by the 
man who seeks to satisfy himself in their realisation. Indeed, 
unless in the stage of conformity to conventional morality 
the principle of obedience is some feeling (though not a clear 
conception) of what should be, of the desirable as distinct 
from the desired, — if it is merely fear of pain or hope of 
pleasure, — there is no approach to autonomy of the will or 
moral freedom in the conformity. We must not allow the 
doctrine that such freedom consists in a determination of the 
will by reason, and the recognition of the truth that the 
requirements of conventional morality are a product of 
reason as operating in individuals of the past, to mislead us 
into supposing that there is any moral freedom, or anything 
of intrinsic value, in the life of conventional morality as 
governed by ' interested motives,' by the desire, directly or 
indirectly, to obtain pleasure. There can be no real deter- 
mination of the will by reason unless both reason and will are 
operating in one and the same person. A will is not really 
anything except as the will of a person, and, as we have seen, 
a will is not really determinable by anything foreign to itself : 
it is only determinable by an object which the person willing 
makes his own. As little is reason really anything apart 
from a self-conscious subject, or as other than an idea of per- 
fection to be realised in and by such a subject. The de- 
termination of will by reason, then, which constitutes moral 
freedom or autonomy, must mean its determination by an 
object which a person willing, in virtue of his reason, presents 


to himself, that object consisting in the realisation of an 
idea of perfection in and by himself. Kant's view that the 
action which is merely ' pflichtmassig,' not done 'aus 
Pflicht,' is of no moral value in itself, whatever may be its 
possible value as a means to the production of the will which 
does act ' aus Pflicht,' is once for all true, though he may 
have taken too narrow a view of the conditions of actions 
done ' aus Pflicht,' especially in supposing (as he seems to 
do) that it is necessary to them to be done painfully. There 
is no determination of will by reason, no moral freedom, in 
conformity of action to rules of which the establishment is 
due to the operation of reason or the idea of pei^fection in 
men, unless the principle of conformity in the persons con- 
forming is that idea itself in some form or other. 


Note of the Editor. 

These lectures, which are partly critical and partly expository, treat of 
the moral grounds upon which the state is hased and upon which ohedience 
to the law of the state is justified. They were delivered in 1879-80, 
following upon the course from which the discussion of Kant's moral 
theory in this volume is taken. The two courses are directly connected, 
civil institutions being throughout regarded as the external expression of the 
moral progress of mankind, and as supplying the material through which 
the idea of perfection must he realised. 

As is implied in section 5, the inquiry into the nature of political obli- 
gation forms part of a wider inquiry into the concrete forms of morality in 
general, ' the detail of goodness. 5 The lecturer had intended to complete 
the course by a consideration of 'social virtues' and 'moral sentiments' ; but 
this intention was not carried out. (See section 251.) 




1. The subject of this course of lectures is the principles 
of political obligation ; and that term is intended to include 
the obligation of the subject towards the sovereign, the 
obligation of the citizen towards the state, and the obligation 
of individuals to each other as enforced by a political superior. 
My purpose is to consider the moral function or object 
served by law, or by the system, of rights and obligations 
which the state enforces, and in so doing to discover the true 
ground or justification for obedience to law. My plan will 
be (1) to state in outline what I consider the true function of 
law to be, this being at the same time the true ground of our 
moral duty to obey the law ; and throughout I distinguish 
moral duty from legal obligation ; (2) to examine the chief 
doctrines of political obligation that have been current in 
modern Europe, and by criticising them to bring out more 
clearly the main points of a truer doctrine ; (3) to consider in 
detail the chief rights and obligations enforced in civilised 
states, inquiring what is their justification, and what is 
the ground for respecting them on the principle stated. 

2. In previous lectures I have explained what I under- 
stand moral goodness to be, and how it is possible that there 
should be such a thing ; in other words, what are the condi- 
tions on the part of reason and will which are implied in our 
being able to conceive moral goodness as an object to be aimed 
at, and to give some partial reality to the conception. Our 
results on this question may be briefly stated as follows. 

The highest moral goodness we found was an attribute 
of chaiacter, in so far as it issued in acts done for the sake 


of their goodness, not for the sake of any pleasure or any 
satisfaction of desire which they bring to the agent. But 
it is impossible that an action should be done for the sake 
of its goodness, unless it has been previously contemplated 
as good for some other reason than that which consists in 
its being done for the sake of its goodness. It must have 
been done, or conceived as possible to be done, and have 
been accounted good, irrespectively of the being done from 
this which we ultimately come to regard as the highest 
motive. In other words, a prior morality, founded upon 
interests which are other than the pure interest in being- 
good, and governed by rules of conduct relative to a standard 
of goodness other than that which makes it depend on this 
interest, is the condition of there coming to be a character 
governed by interest in an ideal of goodness. Otherwise 
this ideal would be an empty one ; it would be impossible to 
say what the good actions were, that were to be done for 
the sake of their goodness ; and the interest in this ideal 
would be impossible, since it would be an interest without 
an object. 

3. When, however, morality of the latter kind has come 
to be recognised as the highest or the only true morality, 
the prior morality needs to be criticised from the point of 
view thus gained. Those interests, other than the interest 
in being good, which form the motives on the part of the 
individual on which it rests, will not indeed be rejected as 
of no moral value ; for no one can suppose that without 
them, or except as regulating them, the pure interest in 
being good could determine conduct at all. But they will 
be estimated according to their value as leading up to, or 
as capable of becoming elements in, a character in which 
this interest is the governing principle. Again, those rules 
of conduct, according to which the terms right and wrong, 
good and bad, are commonly applied, and which, as was just 
now said, are relative to a standard certainly not founded on 
the conception of the good as consisting in the character 
described, are not indeed to be rejected ; for without them 
there would be nothing to define the duties which the highest 
character is prepared to do for their own sake. But they 
have to be revised according to a method which inquires 
into their rationale or justification, as conditions of approxi- 
mation to the highest character. 


4. Such a criticism of moral interests — of the general 
motives which determine moral conduct and regulate such 
moral approbation or disapprobation as is not based on a 
strict theory of moral good — may be called by the name of 
6 a theory of moral sentiments.' The criticism of recognised 
rules of conduct will fall under two heads, according as 
these rules are embodied in positive law (law of which the 
observance is enforced on the individual by a political 
superior), or only form part of the 'law of opinion' (part of 
what the individual feels to be expected of him by some 
person or persons to whose expectations he ought to con- 
form) . 

5. Moral interests are so greatly dependent on generally 
recognised rules of conduct that the criticism of the latter 
should come first. The law of opinion, again, in so many 
ways presupposes a social fabric supported by c positive ? 
law, that we can only fairly take account of it when we have 
considered the moral value and justifiability of the fabric so 
supported. I propose therefore to begin our inquiry into "] 
the detail of goodness — into the particular kinds of conduct 
which the man wishing to do good for the sake of its good- 
ness is entitled to count good — by considering what is of 
permanent moral value in the institutions of civil life, as 
established in Europe ; in what way they have contributed 
and contribute to the possibility of morality in the higher 
sense of the term, and are justified, or have a moral claim 
upon our loyal conformity, in consequence. 

6. The condition of a moral life is the possession of will 
and reason. Will is the capacity in a man of being deter- 
mined to action by the idea of a possible satisfaction of 
himself. An act of will is an action so determined. A 
state of will is the capacity as determined by the particular 
objects in which the man seeks self-satisfaction ; and it 
becomes a character in so far as the self-satisfaction is 
habitually sought in objects of a particular kind. Practical 
reason is the capacity in a man of conceiving the perfection 
of his nature as an object to be attained by action. All 
moral ideas have their origin in reason, i.e. in the idea of a 
possible self-perfection to be attained by the moral agent. 
This does not mean that the moral agent in every stage of 
his progress could state this idea to himself in an abstract 
form, any more than in every stage in the acquisition of 

VOL. II. 7 


knowledge about nature a man can state to himself in an 
abstract form the conception of the unity of nature, which 
yet throughout conditions the acquisition of his knowledge. 
Ideas do not first come into existence, or begin to operate, 
upon the formation of an abstract expression for them. 
This expression is only arrived at upon analysis of a concrete 
experience, which they have rendered possible. Thus we 
only learn to express the idea of self-perfection in that 
abstract form upon an analysis of an experience of self- 
improvement which we have ourselves gone through, and 
which must have been gone through by those with whom 
we are connected by the possession of language and an 
organisation of life, however elementary : but the same 
analvsis shows that the same idea must have been at work 
to make such experience possible. In this idea all particular 
moral ideas — all ideas of particular forms of conduct as 
estimable — originate, though an abstract expression for the 
latter is arrived at much sooner than such an expression 
for the idea in which they originate. They arise, as the 
individual's conception of the society on the well-being of 
which his own depends, and of the constituents of that well- 
being, becomes wider and fuller ; and they are embodied in 
the laws, institutions, and social expectation, which make 
conventional morality. This embodiment, again, constitutes 
the moral progress of mankind. This progress, however, is 
only a moral progress in so far as it tends to bring about 
the harmony of will and reason, in the only form in which 
it can really exist, viz. in the characters of persons. And 
this result is actually achieved, in so far as upon habits 
disciplined by conformity to conventional morality there 
supervenes an intelligent interest in some of the objects 
contributory to human perfection, which that conventional 
morality subserves, and in so far as that interest becomes 
the dominant interest of the character. 

7. The value then of the institutions of civil life lies in 
their operation as giving reality to these capacities of will 
and reason, and enabling them to be really exercised. In 
their general effect, apart from particular aberrations, they 
render it possible for a man to be freely determined by the 
idea of a possible satisfaction of himself, instead of being 
driven this way and that by external forces, and thus they 
give reality to the capacity called will : and they enable 


him to realise his reason, i.e. his idea of self-perfection, by 
acting as a member of a social organisation in which each 
contributes to the better-being of all the rest. So far as 
they do in fact thus operate they are morally justified, and 
may be said to correspond to the c law of nature,' the jus 
naturce, according to the only sense m which that phrase 
can be intelligibly used. 

8. There has been much controversy as to what the jus 
naturce (' Naturrecht ') really is, or whether there is such a 
thing at all. And the controversy, when it comes to be 
dealt with in English, is further embarrassed by the fact that 
we have no one term to represent the full meaning of * jus ' 
or 'Kecht,' as a system of correlative rights and obligations, 
actually enforced or that should be enforced by law. But 
the essential questions are : (1) whether we are entitled to 
distinguish the rights and obligations which are anywhere 
actually enforced by law from rights and obligations which 
really exist though not enforced; and (2), if we are entitled 
to do so, what is to be our criterion of rights and obligations 
which are really valid, in distinction from those that are 
actually enforced. 

9. No one would seriously maintain that the system of 
rights and obligations, as it is anywhere enforced by law 
— the 'jus ' or ' Recht ' of any nation — is all that it ought to 
be. Even Hobbes holds that a law, though it cannot be 
unjust, may be pernicious. But there has been much 
objection to the admission of natural rights and obligations. 
At any rate the phrase is liable to misinterpretation. It 
may be taken to imply that rights and obligations can exist 
in a f state of nature ' — a state in which every individual is 
free to do as he likes — ; that legal rights and obligations 
derive their authority from a voluntary act by which indivi- 
duals contracted themselves out of this state ; and that the 
individual retains from the state of nature certain rights 
with which no legal obligations ought to conflict. Such a 
doctrine is generally admitted to be untenable ; but it does 
not follow from this that there is not a true and important 
sense in which natural rights and obligations exist, — the same 
sense as that in which duties may be said to exist though 
unfulfilled. There is a system of rights and obligations which 
should be maintained by law, whether it is so or not, and 
which may properly be called ' natural ' ; not in the sense in 

z 2 


which the term s natural ' would imply that such a system 
ever did exist or could exist independently of force exercised 
by society over individuals, but ' natural ' because necessary to 
the end which it is the vocation of human society to realise. 

10. The 'jus naturse,' thus understood, is at once distin- 
guished from the sphere of moral duty, and relative to it. 
It is distinguished from it because admitting of enforcement 
by law. Moral duties do not admit of being so enforced. 
The question sometimes put, whether moral duties should 
be enforced by law, is really an unmeaning one ; for they 
simply cannot be enforced. They are duties to act, it is 
true, and an act can be enforced : but they are duties to act 
from certain dispositions and with certain motives, and these 
cannot be enforced. Nay, the enforcement of an outward 
act, the moral character of which depends on a certain 
motive and disposition, may often contribute to render that 
motive and disposition impossible : and from this fact arises 
a limitation to the proper province of law in enforcing 
acts, which will have to be further considered below. When 
obligations then are spoken of in this connection, as part of 
the 'jus naturse ' correlative to rights, they must always be 
understood not as moral duties, not as relative to states of 
will, but as relative to outward acts, of which the perform- 
ance or omission can and should be enforced. There is a 
moral duty to discharge such obligations, and to do so in a 
certain spirit, but the obligation is such as that with which 
law has to do or may have to do, is relative to an outward 
act merely, and does not amount to a moral duty. There is 
a moral duty in regard to obligations, but there can be no 
obligation in regard to moral duties. Thus the ' jus natura? ' 
— the system of rights and obligations, as it should become 
no less than as it actually is maintained — is distinct from 
morality in the proper sense. But it is relative to it. This 
is implied in saying that there is a moral duty in regard to 
actual obligations, as well as in speaking of the system of 
rights and obligations as it should become. If such lan- 
guage is justifiable, there must be a moral ground both for 
conforming to, and for seeking to develope and improve, 
established ' Eecht ' ; a moral ground which can only lie in 
the moral end served by that established system. 

11. Thus we begin the ethical criticism of law with two 
principles : — (I) that nothing but external acts can be 


matter of ' obligation ' (in the restricted sense) ; and (2) 
that, in regard to that which can be made matter of obliga- 
tion, the question what should be made matter of obligation 
— the question how far rights and obligations, as actually 
established by law, correspond to the true ' jus natnrse ' — 
mast be considered with reference to the moral end, as 
serving which alone law and the obligations imposed by law 
have their value. 1 

12. Before proceeding, some remarks have to be made as 
to what is implied in these principles, (a) Does the law, or 
is it possible that it should, confine its view to external acts ? 
What exactly is meant by an external act ? In the case of 
obligations which I am legally punishable for disregarding, 
the law, in deciding whether punishment is or is not due, 
takes account of much beside the external act ; and this im- 
plies that much beside external action is involved in legal 
obligation. In the case where the person or property of 
another is damaged by me, the law does not inquire merely 
whether the act of damage was done, and done by means of 
my bodily members, but whether it was done intentionally ; 
and if not done with the direct intention of inflicting the 
damage, whether the damage arose in a manner that might 
have been foreseen out of something which I did intend to 
do : whether, again, if it was done quite accidentally the 

1 There are two definitions of ' Reeht' ' Eight is that which is really necessary 

or 'jus naturaV quoted by Ulrici to the maintenance of the material con- 

(Naturrecht, p. 219), which embody the ditions essential to the existence and 

truths conveyed in these statements. perfection of human personality.' Cf. 

(]) Krause defines 'Recht' as 'das Trendelenburg, Natur recht, § 46. 'Das 

organische Ganze der ausseren Bedin- Recht ist im sittlichen Ganzen der In- 

gungen des Vernunftlebens,' ' the organic begriff derjenigen allgemeinen Bestim- 

whole of the outward conditions neces- mungen des Handelns, durch welcho 

sary to the rational life.' (2) Henrici es geschieht dass das sittliche Ganze 

says that ' Recht ' is ' was der Idee der und seine Gliederung sich erhalten uud 

Unverletzbarkeit der materiellen we- weiter bilden kann.' Afterwards he 

sentlichenBedingungendesmoralischen emphasises the words 'des Handelns,' 

Menschentbums, d. h. der menschlichen and adds : ' Zwar kann das Hand el n 

Personlichkeit nach ihrer Existenz und nicht ohne den Willen gedacht werden, 

ihrer Vervollkommnung, oder der un- der zum Grundeliegt: aber die Recht- 

verausserlichen Menschengiiter im bestimmungen sind nicht Bestimmungen 

ausserlichen Verkehr entspricht': i.e. des Willens als solchen, wasdem innern 

'Right is what' (or, 'that is properly Gebiet, der Ethik der Gesinnung, 

matter of legal obligation which ') 'in anheimfallen wiirde. Der Wille der 

the outward intercourse of men corre- nicht Handlung wird entzieht sich dcm 

sponds to the idea of the inviolability Recht. Wenn das Recht Sehuld und 

of the essential material conditions of Verselien, dolus und cu?pa, in sein 

a moral humanity, i.e. of the human Bereich zieht, so sind sie als innere aber 

personality in respect of its existence charakteristische Beschaffenheiten des 

and its perfection;' or, more simply, Handelns anzusehen.' 


accident was due to culpable negligence. This, however, does 
not show that the law can enforce or prevent anything but 
external action, but only that it is action which it seeks to 
\ enforce or prevent, for without intention there is no action. 
We talk indeed of a man acting against his will, but if this 
means acting against intention it is what it is impossible 
■to do. What I call an act done against my will is either (1 ) 
an act done by someone else using my body, through superior 
force, as a means : in which case there is an act, but it is not 
mine (e.g. if another uses my hand to pull the trigger of a 
gun by which someone is shot) ; or (2) a natural event in 
which my limbs are affected in a certain way which causes 
certain results to another person (e.g. if the rolling of a ship 
throws me against another person who is thus thrown into 
the water) ; or (3) an act which I do under the influence of 
some strong inducement (e.g. the fear of death), but which is 
contrary to some strong wish. In this case the act is mine, 
but mine because I intend it ; because it is not against my 
will as = intention. In saying, then, that the proper, because 
the only possible, function of law is to enforce the perform- 
ance of or abstinence from external actions, it is implied that 
its function is to produce or prevent certain intentions, for 
without intention on the part of someone there is no act. 

13. But if an act necessarily includes intention, what is 
the nature of the restriction implied in calling it external ? 
An external action is a determination of will as exhibited in 
certain motions of the bodily members which produce certain 
effects in the material world ; not a determination of the 
will as arising from certain motives and a certain disposition. 
All that the law can do is to enjoin or forbid determinations 
of will as exhibited in such motions, &c. It does indeed pre- 
sent a motive, for it enforces its injunctions and prohibitions 
primarily by fear — by its threat of certain consequences if its 
commands are disobeyed. This enforcement is not an exer- 
cise of physical force in the strict sense, for in this sense no 
force can produce an action, since it cannot produce a deter- 
mination of will ; and the only way in which the law or its 
administrators employ such force is not in the production but 
in the prevention of action (as when a criminal is locked up 
or the police prevent mischievous persons from assaulting 
us or breaking into our houses). But though, in enforcing 
its commands by threats, the law is presenting a motive, and 


thus, according" to our distinction, affecting action on its 
inner side, it does this solely for the sake of the external act. 
It does not regard the relation of the act to the motive fear 
as of any intrinsic importance. If the action is performed 
without this motive ever coming into play under the influence 
of what the moralist counts higher motives, the purpose of 
the law is equally satisfied. Indeed, it is always understood 
that its purpose is most thoroughly served when the threat 
of pains and penalties has ceased to be necessary, and the 
obligations correlative to the relations of individuals and of 
societies are fulfilled from other motives. Its business is to 
maintain certain conditions of life — to see that certain actions 
are done which are necessary to the maintenance of those 
conditions, others omitted which would interfere witb them. 
It has nothing to do with the motive of the actions or 
omissions, on which, however, the moral value of them 

14. It appears, then, that legal obligations — obligations 
which can possibly form the subject of positive law — can only 
be obligations to do or abstain from certain acts, not duties 
of acting from certain motives, or with a certain disposition. 
It is not a question whether the law should or should not 
oblige to anything but performance of outward acts. It 
simply cannot oblige to anything else, because the only 
means at its command for obtaining the fulfilment of obli- 
gations are (1) threats of pain and offers of reward, by means 
of which it is possible indeed to secure the general perform- 
ance of certain acts, but not their performance from the 
motive even of fear of the pain threatened or hope of the 
reward offered, much less from any higher motive ; (2) the 
employment of physical force, (a) in restraining men dis- 
posed to violate obligations, (b) in forcibly applying the 
labour or the property of those who violate obligations to 
make good the breach, so far as is possible ; (as, e.g. when 
the magistrate forestalls part of a man's wages to provide for 
a wife whom he has deserted, or when the property of a 
debtor is seized for the benefit of his creditors.) 

15. Only outward acts, then, can be matter of legal obli- 
gation ; but what sort of outward acts should be matter of 
legal obligation ? The answer to this question arises out of 
the above consideration of the means which law employs to 
obtain the fulfilment of obligations, combined with the view 


of law as relative to a moral end, i.e. the formation of a 
society of persons, acting from a certain disposition, from 
interest in the society as such. Those acts only should be matter 
of legal injunction or prohibition of which the performance 
or omission, irrespectively of the motive from which it pro- 
ceeds, is so necessary to the existence of a society in which the 
moral end stated can be realised, that it is better for them to 
be done or omitted from that unworthy moHve which consists 
in fear or hope of legal consequences than not to be done at all. 

16. We distinguish, then, the system of rights actually 
maintained and obligations actually enforced by legal 
sanctions ('Recht' or 'jus') from the system of relations 
and obligations which should be maintained by such sanctions 
(' Naturrecht ') ; and we hold that those actions or omissions 
should be made obligations which, when made obligations, 
serve a certain moral end ; that this end is the ground or 
justification or rationale of legal obligation; and that thus 
we obtain a general rule, of both positive and negative ap- 
plication, in regard to the proper matter or content of legal 
obligation. For since the end consists in action proceeding 
from a certain disposition, and since action done from appre- 
hension of legal consequences does not proceed from that 
disposition, no action should be enjoined or prohibited by 
law of which the injunction or prohibition interferes with 
actions proceeding from that disposition, and every action 
should be so enjoined of which the performance is found to 
produce conditions favourable to action proceeding from that 
disposition, and of which the legal injunction does not inter- 
fere with such action. 

17. Does this general rule give any real guidance in the 
difficulties which practically arise in regard to the province 
of law — as to what should be required by law, and what left 
to the inclination of individuals ? What cases are there or 
have there been of enactments which on this principle we 
can pronounce wrong? Have attempts ever been made by 
law to enforce acts as virtuous which lose their virtue when 
done under fear of legal penalties P It would be difficult, no 
doubt, to find instances of attempts to enforce bylaw actions 
of which we should say that the value lies in the disposition 
from which they are done, actions, e.g. of disinterested 
kindness, because the clear conception of virtue as de- 
pending not on outward results, but on disposition, is but 


slowly arrived at, and has never been reflected in law. But 
without any strictly moral object at all, laws have been made 
which check the development of the moral disposition. 
This has been done (a) by legal requirements of religious 
observance and profession of belief, which have tended to 
vitiate the religious source of morality ; (b) by prohibitions 
and restraints, unnecessary, or which have ceased to be 
necessary, for maintaining the social conditions of the moral 
life, and which interfere with the growth of self-reliance, 
with the formation of a manly conscience and sense of moral 
dignity, — in short, with the moral autonomy which is the 
condition of the highest goodness ; (c) by legal institutions 
which take away the occasion for the exercise of certain 
moral virtues (e.g. the Poor-law which takes away the oc- 
casion for the exercise of parental forethought, filial reverence, 
and neighbourly kindness). 

18. Laws of this kind have often been objected to on the 
strength of a one-sided view of the function of laws ; the 
view, viz. that its only business is to prevent interference 
with the liberty of the individual. And this view has 
gained undue favour on account of the real reforms to which 
it has led. The laws which it has helped to get rid of were 
really mischievous, but mischievous for further reasons than 
those conceived of by the supporters of this theory. Having 
done its work, the theory now tends to become obstructive, 
because in fact advancing civilisation brings with it more 
and more interference with the liberty of the individual to 
do as he likes, and this theory affords a reason for resisting 
all positive reforms, all reforms which involve an action of 
the state in the way of promoting conditions favourable to 
moral life. It is one thing to say that the state in promot- 
ing these conditions must take care not to defeat its true 
end by narrowing the region within which the spontaneity 
and disinterestedness of true morality can have play; 
another thing to say that it has no moral end to serve at all, 
and that it goes beyond its province when it seeks to do 
more than secure the individual from violent interference by 
other individuals. The true ground of objection to ' paternal 
government ' is not that it violates the ' laissez faire ' 
principle and conceives that its office is to make people 
good, to promote morality, but that it rests on a misconcep- 
tion of morality. The real function of government being to 



maintain conditions of life in which morality shall be 
possible, and morality consisting in the disinterested per- 
formance of self-imposed duties, 'paternal government' does 
its best to make it impossible by narrowing the room for 
the self-imposition of duties and for the play of disinterested 

19. The question before us, then, is, In what ways and 
how far do the main obligations enforced and rights main- 
tained by law in all civilised societies contribute to the moral 
end described ; viz. to establish those conditions of life in 
which a true i.e. a disinterested or unselfish morality shall 
be possible ? The answer to this question will be a theory of 
the 'jus naturse' ; i.e. it will explain how far positive law is 
what it should be, and what is the ground of the duty to 
obey it ; in other words, of political obligation. There are 
two things from which such a theory must be distinguished. 
(1) It is not an inquiry into the process by which actual 
law came to be what it is ; nor (2) is it an inquiry how far 
actual law corresponds to and is derived from the exercise 
of certain original or natural rights. (1) It is not the 
former, because the process by which the law of any nation 
and the law in which civilised nations agree has come to 
be what it is, has not been determined by reference to that 
end to which we hold that law ought to be directed and 
by reference to which we criticise it. That is to say, the 
process has not been determined by any such conscious 
reference on the part of the agents in the process. No 
doubt a desire for social good as distinct from private 
pleasure, for what is good on the whole as distinct from 
what is good for the moment, has been a necessary condition 
of it; but (a), as an agent in the development of law, this 
has not reached the form of a conception of moral good 
according to that definition of it by which the value of law 
is to be estimated ; and (b) in bringing law to its present 
state it has been indistinguishably blended with purely 
selfish passions and with the simple struggle for existence. 

20. (2) A true theory of ' jus naturae,' a rationale of law 
or ideal of what it should be, is not to be had by inquiring 
how far actual law corresponds to, and is derived from, the 
exercise of certain original or natural rights, if that is taken 
to mean that we know, or can ascertain, what rights are 
natural on grounds distinct from those on which we deter- 


mine what laws are justifiable, and that then we can proceed 
to ascertain what laws are justifiable by deduction from 
such rights. ' Natural rights,' so far as there are such things, 
are themselves relative to the moral end to which perfect 
law is relative. A law is not good because it enforces 
' natural rights,' but because it contributes to the realisation 
of a certain end. We only discover what rights are natural 
by considering what powers must be secured to a man in 
order to the attainment of this end. These powers a perfect 
law will secure to their full extent. Thus the consideration 
of what rights are ' natural ' (in the only legitimate sense) 
and the consideration what laws are justifiable form one and 
the same process, each presupposing a conception of the 
moral vocation of man. 

21. The doctrine here asserted, that all rights are relative 
to moral ends or duties, must not be confused with the 
ordinary statement that every right implies a duty, or that 
rights and duties are correlative. This of course is true in 
the sense that possession of a right by any person both 
implies an obligation on the part of someone else, and is 
conditional upon the recognition of certain obligations on 
the part of the person possessing it. But what is meant is 
something different, viz. that the claim or right of the 
individual to have certain powers secured to him by society, 
and the counter-claim of society to exercise certain powers 
over the individual, alike rest on the fact that these powers 
are necessary to the fulfilment of man's vocation as a moral 
being, to an effectual self-devotion to the work of developing 
the perfect character in himself and others. 

22. This, however, is not the ground on which the claim 
in question has generally been asserted. Apart from the 
utilitarian theory, which first began to be applied politically 
by Hume, the ordinary way of justifying the civil rights of 
individuals (i.e. the powers secured to them by law as 
against each other), as well as the rights of the state against 
individuals (i.e. the powers which, with the general approval 
of society, it exercises against them), has been to deduce 
them from certain supposed prior rights, called natural rights. 
In the exercise of these natural rights, it has been supposed, 
men with a view to their general interest established political 
society. From that establishment is derived both the system 
of rights and obligations maintained by law as between 


man and man, and the right of the state to the sub- 
mission of its subjects. If the question, then, is raised, 
why I ought to respect the legal rights of my neighbours, 
to pay taxes, or have my children vaccinated, serve in the 
army if the state requires it, and generally submit to the 
law, the answer according to this theory will be that if I 
fail to do so, I shall directly or indirectly be violating the 
natural rights of other men ; directly in those cases where 
the legal rights of my neighbours are also natural rights, as 
they very well may be (e.g. rights of liberty or personal 
safety) ; indirectly where this is not the case, because, 
although the rights of the state itself are not natural, and 
many rights exercised by individuals would not only not be 
secured but would not exist at all but for legal enactment, 
yet the state itself results from a covenant which originally, 
in the exercise of their natural rights, men made with each 
other, and to which all born under the state and sharing 
the advantages derived from it must be considered parties. 
There is a natural right, therefore, on the part of each 
member of a state to have this compact observed, with a cor- 
responding obligation to observe it ; and this natural right 
of all is violated by any individual who refuses to obey the 
law of the state or to respect the rights, not in themselves 
natural, which the state confers on individuals. 

23. This, on the whole, was the form in which the ground 
of political obligation, the justification of established rights, 
was presented throughout the seventeenth century, and in 
the eighteenth till the rise of the ' utilitarian ' theory of 
obligation. Special adaptations of it were made by Hobbes 
and others. In Hobbes, perhaps (of whom more later), may 
be found an effort to fit an anticipation of the utilitarian 
theory of political obligation into the received theory which 
traced political obligation, by means of the supposition of a 
primitive contract, to an origin in natural right. But in 
him as much as anyone the language and framework of 
the theory of compact is retained, even if an alien doctrine 
may be read between the lines. Of the utilitarian theory of 
political obligation more shall be said later. It may be pre- 
sented in a form in which it would scarcely be distinguishable 
from the doctrine just now stated, the doctrine, viz. that 
the ground of political obligation, the reason why certain 
powers should be recognised as belonging to the state and 


certain other powers as secured by the state to individuals, 
lies in the fact that these powers are necessary to the fulfil- 
ment of man's vocation as a moral being, to an effectual self- 
devotion to the work of developing the perfect character in 
himself and others. Utilitarianism proper, however, recog- 
nises no vocation of man but the attainment of pleasure and 
avoidance of pain. The only reason why civil rights should 
be respected — the only justification of them — according to it, 
would be that more pleasure is attained or pain avoided by 
the general respect for them ; the ground of our conscious- 
ness that we ought to respect them, in other words their 
ultimate sanction, is the fear of what the consequences would 
be if we did not. This theory and that which I deem true 
have one negative point in common. They do not seek the 
ground of actual rights in a prior natural right, but in an end 
to which the maintenance of the rights contributes. They 
avoid the mistake of identifying the inquiry into the ultimate 
justifiability of actual rights with the question whether there 
is a prior right to the possession of them. The right to the 
possession of them, if properly so called, would not be a mere 
power, but a power recognised by a society as one which 
should exist. This recognition of a power, in some way or 
other, as that which should be, is always necessary to render 
it a right. Therefore when we had shown that the rights 
exercised in political society were derived from prior ' natural ' 
rights, a question would still remain as to the ground of those 
natural rights. We should have to ask why certain powers 
were recognised as powers which should be exercised, and 
thus became these natural rights. 

24. Thus, though it may be possible and useful to show 
how the more seemingly artificial rights are derived from 
rights more simple and elementary, how the rights esta- 
blished by law in a political society are derived from rights 
that may be called natural, not in the sense of being prior to 
society, but in the sense of being prior to the existence of 
a society governed by written law or a recognised sovereign, 
still such derivation is no justification of them. It is no 
answer to the question why they should be respected ; because 
this question remains to be asked in regard to the most 
primitive rights themselves. Political or civil rights, then, 
are not to be explained by derivation from natural rights, 
but in regard to both political and natural rights, in any sense 


in which there can be truly said to be natural rights, the ques- 
tion has to be asked, how it is that certain powers are recog- 
nised by men in their intercourse with each other as powers 
that should be exercised, or of which the possible exercise 
should be secured. 

25. I have tried to show in lectures on morals that the 
conception expressed by the ' should be ' is not identical 
with the conception of a right possessed by some man or 
men, but one from which the latter conception is derived. 
It is, or implies on the part of whoever is capable of it, the 
conception of an ideal, unattained, condition of himself, as 
an absolute end. Without this conception the recognition 
of a power as a right would be impossible. A power on the 
part of anyone is so recognised by others, as one which 
should be exercised, when these others regard it as in some 
way a means to that ideal good of themselves which they 
alike conceive : and the possessor of the power comes to 
regard it as a right through consciousness of its being thus 
recognised as contributory to a good in which he too is 
interested. No one therefore^ can have a right except (1) as 
a member of a~~society, and (2) of a society in which some 
common good is recognised by the members of the society 
as their own ideal good, as that which should be for each 
of them. The capacity for being determined by a good so 
recognised is what constitutes personality in the ethical 
sense ; and for this reason there is truth in saying that only 
among persons, in the ethical sense, can there come to be 
rights ; (which is quite compatible with the fact that the 
logical disentanglement of the conception of rights precedes 
that of the conception of the legal person ; and that the 
conception of the moral person, in its abstract and logical 
form, is not arrived at till after that of the legal person). 

Conversely, everyone capable of being determined by the 
conception of a common good as his own ideal good, as that 
which unconditionally should be (of being in that sense 
an end to himself), in other words, every moral person, is 
capable of rights ; i.e. of bearing his part in a society in 
which the free exercise of his powers is secured to each 
member through the recognition by each of the others as 
entitled to the same freedom with himself. To say that he 
is capable of rights, is to say that he ought to have them, in 
that sense of ' ought ' in which it expresses the relation of 


man to an end conceived as absolutely good, to an end 
which, whether desired or no, is conceived as intrinsically 
desirable. The moral capacity implies a consciousness on 
the part of the subject of the capacity that its realisation is 
an end desirable in itself, and rights are the condition of 
realising it. Only through the possession of rights can the 
power of the individual freely to make a common good his 
own have reality given to it. Rights are what may be called 
the negative realisation of this power. That is, they realise 
it in the sense of providing for its free exercise, of securing 
the treatment of one man by another as equally free with 
himself, but they do not realise it positively, because their 
possession does not imply that in any active way the indivi- 
dual makes a common good his own. The possession of 
them, however, is the condition of this positive realisation 
of the moral capacity, and they ought to be possessed because 
this end (in the sense explained) ought to be attained. 

26. Hence on the part of every person (< person ' in the 
moral sense explained) the claim, more or less articulate and 
reflected on, to rights on his own part is co-ordinate with 
hj^j^cogmtionofrights on the part of others. Thejca_pacity 
to conceive a common goocTas one's own, and to regulate the 
exerciseoTone's powers by reference to a~good which others 
recognise, carries with it the consciousness that powers 
should be so exercised ; which means that there should be 
rights, that powers should be regulated by mutual recogni- 
tion. There ought to be rights, because the moral person- 
ality , — the capacity on the part of an i ndividual for mflTrjn g 
a common good his own, — ought to be developed ; and it is 
developed through, rights ; i.e. through the recognition by 
members of a society of powers in each other Contributory 
to a common good, and the regulation of those powers by 
that recognition. 

27. In saying that only among < persons ' can there come 
to be rights, and that every ' person ' should have rights, I 
have been careful to explain that I use ' person ' in the 
moral, not merely in the legal, sense. In dealing, then, with 
such phrases as ' jura personarum ' and ' personal rights,' we 
must keep in view the difference between the legal and 
ethical sense of the proposition that all rights are personal, 
or subsist as between persons. In the legal sense, so far as 
it is true, — and it is so only if ' person ' is used in the sense 


i of Soman law, — it is an identical proposition. A person 
means a subject of rights and nothing more. Legal person- 
ality is derived from the possession of right, not vice versa. 
Like other identical propositions, its use is to bring out and 
emphasise in the predicate what is included in the under- 
stood connotation of the subject; to remind us that when we 
speak of rights we imply the existence of parties, in English 
phraseology, capable of suing and being sued. In the ethical 
sense, it means that rights are derived from the possession 
of personality as = a rational will (i.e. the capacity which 
man possesses of being determined to action by the concep- 
tion of such a perfection of his being as involves the perfec- 
tion of a society in which he lives), in the sense (a) that 
only among beings possessed of rational will can there come 
to be rights, (b) that they fulfil their idea, or are justifiable, 
or such rights as should be rights, only as contributing to 
the realisation of a rational will. It is important to bear 
this distinction in mind in order that the proposition in its 
ethical sense, which can stand on its own merits, may not 
derive apparent confirmation from a juristic truism. 

28. The moral idea of personality is constantly tending to 
affect the legal conception of the relation between rights and 
persons. Thus the * jura personarum,' which properly = 
either rights arising out of e status,' or rights which not only 
(like all rights) reside in someone having a legal status and 
are available against others having a legal status, but are 
exercised over, or in respect of, someone possessed of such 
status (e.g. a wife or a servant), come to be understood as 
rights derived from the human personality or belonging to 
man as man. It is with some such meaning that English 
writers on law speak of rights to life and liberty as personal 
rights. The expression might seem pleonastic, since no right 
can exist except as belonging to a person in the legal sense. 
They do not use the phrase either pleonastically or in the 
sense of the Roman lawyers' ' jura personarum ' above, but 
in the sense that these rights are immediately derived from, 
or necessarily attach to, the human personality in whatever 
that personality is supposed to consist. There is no doubt, 
however, that historically the conception of the moral person, 
in any abstract form, is not arrived at till after that of the 
legal person has been thus disentangled and formulated ; and 
further that the abstract conception of the legal person, as 


the sustainer of rights, is not arrived at till long after rights 
have been actually recognised and established. But the dis- 
entanglement or abstract formulation of the conception of 
moral personality is quite a different thing from the action 
of the consciousness in which personality consists. 

29. The capacity, then, on the part of the individual of 
conceiving a good as the same for himself and others, and of 
bein g determined to action by that conception , is the foundation 
of rights ; and rights are the condition of that capacity being 
realised. No right is justifiable or should be a right except 
on the ground that directly or indirectly it serves this pur- 
pose. Conversely every power should be a right, i.e. society 
should secure to the individual every power, that is necessary 
for realising this capacity. Claims to such powers as are 
directly necessary to a man's acting as a moral person at all 
— acting under the conception of a good as the same for 
self and others — may be called in a special sense personal 
rights (though they will include more than Stephen includes 
under that designation) ; they may also be called, if we avoid 
misconceptions connected with these terms, ' innate ' or 
' natural ' rights. They are thus distinguished from others 
which are (1) only indirectly necessary to the end stated, or 
(2) are so only under special conditions of society ; as well as 
from claims which rest merely on legal enactment and might 
cease to be enforced without any violation of the 'jus 

30. The objection to calling them ' innate ' or ' natural,' 
when once it is admitted on the one side that rights are not 
arbitrary creations of law or custom but that there are certain 
powers which ought to be secured as rights, on the other 
hand that there are no rights antecedent to society, none 
that men brought with them into a society which they con- 
tracted to form, is mainly one of words. They are ' innate ' 
or ' natural ' in the same sense in which according to Aristotle 
the state is natural ; not in the sense that they actually exist 
when a man is born and that they have actually existed as 
long as the human race, but that they arise out of, and are 
necessary for the fulfilment of, a moral capacity without which 
a man would not be a man. There cannot be innate rights 
in any other sense than that in which there are innate duties 
of which, however, much less has been heard. Because a group 
of beings are capable each of conceiving an absolute good of 

VOL. II. a A 


himself and of conceiving it to be good for himself as identical 
with, and because identical with, the good of the rest of the 
group, there arises for each a consciousness that the common 
good should be the object of action, i.e. a duty, and a claim 
in each to a power of action that shall be at once secured and 
regulated by the consciousness of a common good on the 
part of the rest, i.e. a right. There is no ground for saying 
that the right arises out of a primary human capacity, and is 
t thus ' innate,' which does not apply equally to the duty. 

31. The dissociation of innate rights from innate 
duties has gone along with the delusion that such rights 
existed apart from society. Men were supposed to have 
existed in a state of nature, which was not a state of societ} T , 
but in which certain rights attached to them as individuals, 
and then to have formed societies by contract or covenant. 
Society having been formed, certain other rights arose 
through positive enactment ; but none of these, it was held, 
could interfere with the natural rights which belonged to 
men antecedently to the social contract or survived it. 
f Such a theory can only be stated by an application to an 
| imaginary state of things, prior to the formation of societies 
\ as regulated by custom or law, of terms that have no mean- 
ing except in relation to such societies. ' Natural right,' as 
b= right in a state of nature which is not a state of society, 
j is a contradiction. There can be no right without a con- 
i sciousness of common interest on the part of members of a 
society. Without this there might be certain powers on the 
\ part of individuals, but no recognition of these powers by 
others as powers of which they allow the exercise, nor any 
claim to such recognition ; and without this recognition or 
claim to recognition there can be no right. 



32. Spinoza is aware of this. In the Tractatus Politici, 
II. 4, he says, ' Per jus itaque naturae intelligo . . . ipsam 
naturae potentiam.' . . . 'Quicquid unusquisque homo ex 
legibus suae naturae agit, id summo naturae jure agit, tantum- 
que in naturam habet juris, quantum potentia valet.' If 
only, seeing that the 'jus naturae' was mere 'potentia,' he 
had denied that it was 'jus' at all, he would have been on 
the right track. Instead of that, however, he treats it as 
properly 'jus,' and consistently with this regards all 'jus' 
as mere 'potentia': nor is any 'jus humanum ' according 
to him guided by or the product of reason. It arises, in 
modern phrase, out of the 'struggle for existence.' As 
Spinoza says, ' homines magis caeca cupiditate quam ratione 
ducuntur ; ac proinde hominum natural is potentia sive jus non 
ratione, sed quocumque appetitu quo ad agendum determi- 
nantur, quoque se conservare conantur, definiri debet ' (II. 5). 
The 'jus civile ' is simply the result of the conflict of natural 
powers, which = natural rights, which arises from the effort 
of every man to gratify his passions and ' suum esse conser- 
vare.' Man is simply a ' pars naturae,' the most crafty of the 
animals. ' Quatenus homines ira, invidia aut aliquo odii 
affectu conflictantur, eatenus diverse trahuntur et invicem 
contrarii sunt, et propterea eo plus timendi, quo plus possunt, 
magisque callidi et astuti sunt, quam reliqua animalia ; et 
quia homines ut plurimum his affectibus natura sunt obnoxii, 
sunt ergo homines ex natura hostes ' (II. 14), Universal 
hostility means universal fear, and fear means weakness. It 
follows that in the state of nature there is nothing fit to be 
called ' potentia ' or consequently ' jus ' ; ' atque adeo con- 
cludimus jus naturae vix posse concipi nisi ubi homines jura 
habent communia, qui simul terras, quas habitare et colere 
possunt, sibi vindicare, seseque munire, vimque omnem repel- 

A A 2 


lere et ex communi omnium sententia vivere possunt. Nam 
(per art. 13 hujus cap.) quo plures in unum sic conveniunt, eo 
o nines simul plus juris habent ' ( 15) . The collective body, i.e., 
has more 'jus in naturam,' i.e. 'potentiam,' than any indivi- 
dual could have singly (13). In the advantage of this in- 
creased ' jus in naturam ' the individual shares. On the other 
hand (16), 'Ubi homines jura communia habent omnesque 
una veluti mente ducuntur, certum est (per art. 13 hujus 
cap.) eorum unumquemque tanto minus habere juris, quanto 
reliqui simul ipso potentiores sunt, hoc est, ilium revera jus 
nullum in naturam habere preeter id, quod ipsi commune 
concedit jus. Ceterum quicquid ex communi consensu ipsi 
imperatur, teneri exsequi vel (per art. 4 hujus cap.) jure ad 
id cogi.' This 'jus' by which the individual's actions are 
now to be regulated, is still simply ' potentia.' ' Hoc jus, 
quod multitudinis potentia definitur, imperium appellari 
solet ' (17). It is not to be considered anything different from 
the ' jus naturse.' It is simply the 'naturalis potentia ' of a 
certain number of men combined ; ' multitudinis quse una 
veluti mente ducitur ' (III. 2). Thus in the 'status civilis ' 
the ' jus naturae ' of the individual in one sense disappears, 
in another does not. It disappears in the sense that the 
individual member of the state has no mind to act or power 
to act against the mind of the state. Anyone who had 
such mind or power would not be a member of the state. 
He would be an enemy against whose ' potentia ' the state 
must measure its own. On the other hand, ' in statu civili,' 
just as much as 'in statu naturali,' 'homo ex legibus suse 
naturae agit suseque utilitati consulit ' (3). He exercises his 
' naturalis potentia ' for some natural end of satisfying his 
wants and preserving his life as he did or would do outside 
the ' status civilis.' Only in the ' status civilis ' these motives 
on the part of individuals so far coincide as to form the 
' una veluti mens ' which directs the ' multitudinis potentia.' 
According to this view, any member of a state will have 
just so much 'jus,' i.e. 'potentia,' against other members 
as the state allows him. If he can exercise any 'jus' or 
' potentia ' against another ' ex suo ingenio,' he is so far not 
a member of the state and the state is so far imperfect. If 
he could exercise any ' jus ■ or ' potentia ' against the state 
itself, there would be no state, or, which is the same, the 
state would not be ' sui juris.' 


33. Is there then no limit to the ' jus' which the state 
may exercise ? With Spinoza this is equivalent to the ques- 
tion, is there no limit to the c potentia ' which it can 
exercise ? As to this, he suggests three considerations. 

(1). Its power is weakened by any action against right 
reason, because this must weaken the ' animorum unio ' on 
which it is founded. * Civitatis jus potentia multitudinis, 
quae una veluti mente ducitur, determinatur. At hsec ani- 
morum unio concipi nulla ratione posset, nisi civitas id 
ipsum maxime intendat, quod sana ratio omnibus hominibus 
utile esse docet ' (III. 7) . 

(2). The ' right' or 'power' of the state depends on its 
power of affecting the hopes and fears of individual citizens. 
. . . ' Subditi eatenus non sui, sed civitatis juris sint, qua- 
tenus ejus potentiam seu minas metuunt, vel quatenus 
statum civilem amant (per art. 10 prseced. cap.). Ex quo 
sequitur, quod ea omnia, ad quse agenda nemo prsemiis aut 
minis induci potest, ad jura civitatis non pertineant ' (III. 
8). Whatever cannot be achieved by rewards and threats, is 
beyond the power and therefore beyond the e right ' of the 
state. Examples are given in the same section. 

(3). 'Ad civitatis jus ea minus pertinere, quae plurimi 
indignantur ' (III. 9). Severities of a certain kind lead to 
conspiracies against the state, and thus weaken it. ' Sicut 
unusquisque civis sive homo in statu naturali, sic civitas eo 
minus sui juris est, quo majorem timendi causam habet.' 

Just so far then as there are certain things which the 
state cannot do, or by doing which it lessens its power, so 
far there are things which it has no ' right ' to do. 

34. Spinoza proceeds to consider the relation of states-^ 
or sovereign powers to each other. Here the principle is 
simple. They are to each other as individuals in the state 
of nature, except that they will not be subject to the same 
weaknesses. 'Nam quandoquidem (per art. 2 hujus cap.) 
jus suminso potestatis nihil est prgeter ipsum naturse jus, 
sequitur duo imperia ad invicem sese habere, ut duo homines 
in statu naturali, excepto hoc, quod civitas sibi cavere potest, 
ne ab alia opprimatur, quod homo in statu naturali non 
potest, nimirum qui quotidie somno, ssepe morbo aut animi 
segritudine, et tandem senectute gravatur, et prseter ha)c aliis 
inconimodis est obnoxius, a quibus civitas securam sereddere 
potest' (III. 11). In other words, '. . . duse civitates 


natura hostes sunt. Homines enim in statu naturali hostes 
sunt. Qui igitur jus naturae extra civitatem retinent, hostes 
manent' (III. 13). The 'jura belli' are simply the powers 
of any one state to attack or defend itself against another. 
The ' jura pacis,' on the other hand, do not appertain to any 
single state, but arise out of the agreement of two at least. 
They last as long as the agreement, the * fcedus,' lasts ; and 
this lasts as long as the fear or hope, which led to its being 
made, continues to be shared by the states which made it. 
As soon as this ceases to be the case, the agreement is 
necessarily at an end, ' nee dici potest, quod dolo vel perfidia 
agat, propterea quod fidem solvit, simulatque metus vel spei 
causa sublata est, quia hsec conditio unicuique contrahentium 
sequalis fuit, ut scilicet quae prima extra metum esse potest, 
sui juris esset, eoque ex sui animi sententia uteretur, et prse- 
terea quia nemo in futurum contrahit nisi positis prseceden- 
tibus circumstantiis ' (III. 14). 

35. It would seem to follow from the above that a state 
can do no wrong, in the sense that there are no rights that 
it can violate. The same principle is applicable to it as 
to the individual. ' In statu naturali non dari peccatum, 
vel si quis peccat, is sibi, non alteri peccat: . . . nihil 
absolute naturse jure prohibetur, nisi quod nemo potest ' (II. 
18). A state is to any other state, and to its subjects, as 
one individual to another ' in statu naturali.' A wrong, a 
'peccatum,' consists in a violation by individuals of the 
' commune decretum.' There can be no ' peccare ' on the 
part of the ' commune decretum ' itself. But * non id omne, 
quod jure fieri dicimus, op time fieri affirmamus. Aliud 
namque est agrum jure colere, aliud agrum optime colere ; 
aliud, inquam, est sese jure defendere, conservare, judicium 
ferre, &c, aliud sese optime defendere, conservare, atque 
optimum judicium ferre; et consequenter aliud est jure 
imperare et reipublicso curam habere, aliud optime imperare et 
rempublicam optime gubernare. Postquam itaque de jure 
cujuscumque civitatis in genere egimus, tempus est, ut de 
optimo cujuscumque imperii statu agamus' (V. 1). Hence 
a further consideration ' de optimo cujusque imperii statu.' 
This is guided by reference to the ' finis status civilis,' which 
is 'pax vitseque securitas.' Accordingly that is the best 
government under which men live in harmony, and of which 
the rights are kept inviolate. Where this is not the case. 


the fault lies with the government, not with any ' subditorum 
malitia.' ' Homines enim civiles non nascuntur, sed fiunt. 
Hominum prseterea naturales affectus ubique iidem sunt' 
(V. 2). 

The end is not fully attained where men are merely kept 
in order by fear. Such a state of things is not peace but 
merely absence of war. ' Pax enim non belli privatio, sed 
virtus est, quse ex animi fortitudine oritur } ; est namque 
obsequium constans voluntas id exsequendi, quod ex communi 
civitatis decreto fieri debet ' (V. 4) . 

The 'peace,' then, which it is the end of the state to 
obtain, consists in rational virtue; in a common mind, 
governed by desire on the part of each individual for perfec- 
tion of being in himself and others. The harmony of life, too, 
which is another way of expressing its object, is to be under- 
stood in an equally high sense. The life spoken of is one 
< quse maxim e ratione, vera mentis virtute et vita, definitur.' 

The * imperium ' which is to contribute to this end must 
clearly be one ' quod multitudo libera instituit, non autem 
id, quod in multitudinem jure belli acquiritur.' Between 
the two forms of 'imperium' there may be no essential 
difference in respect of the 'jus ' wh